Proceedings e report 90 - Firenze University Press

Proceedings e report 90

ECOS 2012 The 25th International Conference on Efficiency, Cost, Optimization and Simulation of Energy Conversion Systems and Processes (Perugia, June 26th-June 29th, 2012)

edited by Umberto Desideri, Giampaolo Manfrida, Enrico Sciubba

firenze university press

2012

ECOS 2012 : the 25th International Conference on Efficiency, Cost, Optimization and Simulation of Energy Conversion Systems and Processes (Perugia, June 26th-June 29th, 2012) / edited by Umberto Desideri, Giampaolo Manfrida, Enrico Sciubba. – Firenze : Firenze University Press, 2012. (Proceedings e report ; 90) http://digital.casalini.it/9788866553229 ISBN 978-88-6655-322-9 (online) Progetto grafico di copertina Alberto Pizarro, Pagina Maestra snc Immagine di copertina: © Kts | Dreamstime.com

Peer Review Process All publications are submitted to an external refereeing process under the responsibility of the FUP Editorial Board and the Scientific Committees of the individual series. The works published in the FUP catalogue are evaluated and approved by the Editorial Board of the publishing house. For a more detailed description of the refereeing process we refer to the official documents published on the website and in the online catalogue of the FUP (http://www.fupress.com). Firenze University Press Editorial Board G. Nigro (Co-ordinator), M.T. Bartoli, M. Boddi, F. Cambi, R. Casalbuoni, C. Ciappei, R. Del Punta, A. Dolfi, V. Fargion, S. Ferrone, M. Garzaniti, P. Guarnieri, G. Mari, M. Marini, M. Verga, A. Zorzi. © 2012 Firenze University Press Università degli Studi di Firenze Firenze University Press Borgo Albizi, 28, 50122 Firenze, Italy http://www.fupress.com/ Printed in Italy

ECOS 2012 The 25th International Conference on Efficiency, Cost, Optimization and Simulation of Energy Conversion Systems and Processes

Perugia, June 26th-June 29th, 2012 Book of Proceedings - Volume VIII Edited by: Umberto Desideri, Università degli Studi di Perugia Giampaolo Manfrida, Università degli Studi di Firenze Enrico Sciubba, Università degli Studi di Roma “Sapienza”

ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY EDITED BY UMBERTO DESIDERI, GIAMPAOLO MANFRIDA, ENRICO SCIUBBA FIRENZE UNIVERSITY PRESS, 2012, ISBN ONLINE : 978-88-6655-322-9

ii

Advisory Committee (Track Organizers) Building, Urban and Complex Energy Systems V. Ismet Ugursal Dalhousie University, Nova Scotia, Canada Combustion, Chemical Reactors, Carbon Capture and Sequestration Giuseppe Girardi ENEA-Casaccia, Italy Energy Systems: Environmental and Sustainability Issues Christos A. Frangopoulos National Technical University of Athens, Greece Exergy Analysis and Second Law Analysis Silvio de Oliveira Junior Polytechnical University of Sao Paulo, Sao Paulo, Brazil Fluid Dynamics and Power Plant Components Sotirios Karellas National Technical University of Athens, Athens, Greece Fuel Cells Umberto Desideri University of Perugia, Perugia, Italy Heat and Mass Transfer Francesco Asdrubali, Cinzia Buratti University of Perugia, Perugia, Italy Industrial Ecology Stefan Goessling-Reisemann University of Bremen, Germany Poster Session Enrico Sciubba University Roma 1 “Sapienza”, Italy Process Integration and Heat Exchanger Networks Francois Marechal EPFL, Lausanne, Switzerland Renewable Energy Conversion Systems David Chiaramonti University of Firenze, Firenze, Italy Simulation of Energy Conversion Systems Marcin Liszka Polytechnica Slaska, Gliwice, Poland System Operation, Control, Diagnosis and Prognosis Vittorio Verda Politecnico di Torino, Italy Thermodynamics A. Özer Arnas United States Military Academy at West Point, U.S.A. Thermo-Economic Analysis and Optimisation Andrea Lazzaretto University of Padova, Padova, Italy Water Desalination and Use of Water Resources Corrado Sommariva ILF Consulting M.E., U.K iii

Scientific Committee Riccardo Basosi, University of Siena, Italy Gino Bella, University of Roma Tor Vergata, Italy Asfaw Beyene, San Diego State University, United States Ryszard Bialecki, Silesian Institute of Tecnology, Poland Gianni Bidini, University of Perugia, Italy Ana M. Blanco-Marigorta, University of Las Palmas de Gran Canaria, Spain Olav Bolland, University of Science and Technology (NTNU), Norway Renè Cornelissen, Cornelissen Consulting, The Netherlands Franco Cotana, University of Perugia, Italy Alexandru Dobrovicescu, Polytechnical University of Bucharest, Romania Gheorghe Dumitrascu, Technical University of Iasi, Romania Brian Elmegaard, Technical University of Denmark , Denmark Daniel Favrat, EPFL, Switzerland Michel Feidt, ENSEM - LEMTA University Henri Poincaré, France Daniele Fiaschi, University of Florence, Italy Marco Frey, Scuola Superiore S. Anna, Italy Richard A Gaggioli, Marquette University, USA Carlo N. Grimaldi, University of Perugia, Italy Simon Harvey, Chalmers University of Technology, Sweden Hasan Heperkan, Yildiz Technical University, Turkey Abel Abel Hernandez-Guerrero, University of Guanajuato, Mexico Jiri Jaromir Klemeš, University of Pannonia, Hungary Zornitza V. Kirova-Yordanova, University "Prof. Assen Zlatarov", Bulgaria Noam Lior, University of Pennsylvania, United States Francesco Martelli, University of Florence, Italy Aristide Massardo, University of Genova, Italy Jim McGovern, Dublin Institute of Technology, Ireland Alberto Mirandola, University of Padova, Italy Michael J. Moran, The Ohio State University, United States Tatiana Morosuk, Technical University of Berlin, Germany Pericles Pilidis, University of Cranfield, United Kingdom Constantine D. Rakopoulos, National Technical University of Athens, Greece Predrag Raskovic, University of Nis, Serbia and Montenegro Mauro Reini, University of Trieste, Italy Gianfranco Rizzo, University of Salerno, Italy Marc A. Rosen, University of Ontario, Canada Luis M. Serra, University of Zaragoza, Spain Gordana Stefanovic, University of Nis, Serbia and Montenegro Andrea Toffolo, Luleå University of Technology, Sweden Wojciech Stanek, Silesian University of Technology, Poland George Tsatsaronis, Technical University Berlin, Germany Antonio Valero, University of Zaragoza, Spain Michael R. von Spakovsky, Virginia Tech, USA Stefano Ubertini, Parthenope University of Naples, Italy Sergio Ulgiati, Parthenope University of Naples, Italy Sergio Usón, Universidad de Zaragoza, Spain Roman Weber, Clausthal University of Technology, Germany Ryohei Yokoyama, Osaka Prefecture University, Japan Na Zhang, Institute of Engineering Thermophysics, Chinese Academy of Sciences, China iv

v

The 25th ECOS Conference 1987-2012: leaving a mark The introduction to the ECOS series of Conferences states that “ECOS is a series of international conferences that focus on all aspects of Thermal Sciences, with particular emphasis on Thermodynamics and its applications in energy conversion systems and processes”. Well, ECOS is much more than that, and its history proves it! The idea of starting a series of such conferences was put forth at an informal meeting of the Advanced Energy Systems Division of the American Society of Mechanical Engineers (ASME) at the November 1985 Winter Annual Meeting (WAM), in Miami Beach, Florida, then chaired by Richard Gaggioli. The resolution was to organize an annual Symposium on the Analysis and Design of Thermal Systems at each ASME WAM, and to try to involve a larger number of scientists and engineers worldwide by organizing conferences outside of the United States. Besides Rich other participants were Ozer Arnas, Adrian Bejan, Yehia ElSayed, Robert Evans, Francis Huang, Mike Moran, Gordon Reistad, Enrico Sciubba and George Tsatsaronis. Ever since 1985, a Symposium of 8-16 sessions has been organized by the Systems Analysis Technical Committee every year, at the ASME Winter Annual Meeting (now ASME-IMECE). The first overseas conference took place in Rome, twenty-five years ago (in July 1987), with the support of the U.S. National Science Foundation and of the Italian National Research Council. In that occasion, Christos Frangopoulos, Yalcin Gogus, Elias Gyftopoulos, Dominick Sama, Sergio Stecco, Antonio Valero, and many others, already active at the ASME meetings, joined the core-group. The name ECOS was used for the first time in Zaragoza, in 1992: it is an acronym for Efficiency, Cost, Optimization and Simulation (of energy conversion systems and processes), keywords that best describe the contents of the presentations and discussions taking place in these conferences. Some years ago, Christos Frangopoulos inserted in the official website the note that “ècos” (’ ) means “home” in Greek and it ought to be attributed the very same meaning as the prefix “Eco-“ in environmental sciences. The last 25 years have witnessed an almost incredible growth of the ECOS community: more and more Colleagues are actively participating in our meetings, several international Journals routinely publish selected papers from our Proceedings, fruitful interdisciplinary and international cooperation projects have blossomed from our meetings. Meetings that have spanned three continents (Africa and Australia ought to be our next targets, perhaps!) and influenced in a way or another much of modern Engineering Thermodynamics. After 25 years, if we do not want to become embalmed in our own success and lose momentum, it is mandatory to aim our efforts in two directions: first, encourage the participation of younger academicians to our meetings, and second, stimulate creative and useful discussions in our sessions. Looking at this years’ registration roster (250 papers of which 50 authored or co-authored by junior Authors), the first objective seems to have been attained, and thus we have just to continue in that direction; the second one involves allowing space to “voices that sing out of the choir”, fostering new methods and approaches, and establishing or reinforcing connections to other scientific communities. It is important that our technical sessions represent a place of active confrontation, rather than academic “lecturing”. In this spirit, we welcome you in Perugia, and wish you a scientifically stimulating, touristically interesting, and culinarily rewarding experience. In line with our 25 years old scientific excellency and friendship! Umberto Desideri, Giampaolo Manfrida, Enrico Sciubba vi

CONTENT MANAGEMENT The index lists all the papers contained all the eight volumes of the Proceedings of the ECOS 2012 International Conference. Page numbers are listed only for papers within the Volume you are looking at. The ID code allows to trace back the identification number assigned to the paper within the Conference submission, review and track organization processes.

vii

CONTENT VOLUME VIII VIII. 1 ENERGY SYSTEMS: ENVIRONMENTAL AND SUSTAINABILITY ISSUES » A multi-criteria decision analysis tool to support electricity planning …….... (ID 467) Fernando Ribeiro, Paula Ferreira, Madalena Araújo

Pag. 1

» Comparison of sophisticated life cycle impact assessment methods …….... for assessing environmental impacts in a LCA study of electricity production (ID 259) Jens Buchgeister

Pag. 15

» Defossilisation assessment of biodiesel life cycle production using the ExROI indicator (ID 304) Emilio Font de Mora, César Torres, Antonio Valero, David Zambrana

……....

Pag. 27

» Design strategy of geothermal plants for water dominant medium-low …….... temperature reservoirs based on sustainability issues (ID 99) Alessandro Franco, Maurizio Vaccaro

Pag. 38

» Energetic and environmental benefits from waste management: …….... experimental analysis of the sustainable landfill (ID 33) Francesco Di Maria, Alessandro Canovai, Federico Valentini, Alessio Sordi, Caterina Micale

Pag. 50

» Environmental assessment of energy recovery technologies for the …….... treatment and disposal of municipal solid waste using life cycle assessment (LCA): a case study of Brazil (ID 512) Marcio Montagnana Vicente Leme, Mateus Henrique Rocha, Electo Eduardo Silva Lora,Osvaldo José Venturini, Bruno Marciano Lopes, Claudio Homero Ferreira

Pag. 62

» How will renewable power generation be affected by climate change? – …….... The case of a metropolitan region in Northwest Germany (ID 503) Jakob Wachsmuth, Andrew Blohm, Stefan Gößling-Reisemann, Tobias Eickemeier, Rebecca Gasper, Matthias Ruth, Sönke Stührmann

Pag. 71

» Impact of nuclear power plant on Thailand power development plan …….... (ID 474) Raksanai Nidhiritdhikrai, Bundhit Eua-arporn

Pag. 88

» Improving sustainability of maritime transport through utilization of …….... liquefied natural gas (LNG) for propulsion (ID 496) Fabio Burel, Rodolfo Taccani, Nicola Zuliani

Pag. 102

» Life cycle assessment of thin film non conventional photovoltaics: the case of dye sensitized solar cells (ID 471) Maria Laura Parisi, Adalgisa Sinicropi, Riccardo Basosi

……....

Pag. 119

» Low CO2 emission hybrid solar CC power system (ID 175) Yuanyuan Li, Na Zhang, Ruixian Cai

……....

Pag. 133

» Low exergy solutions as a contribution to climate adapted and resilient …….... power supply (ID 489) Stefan Goessling-Reisemann, Thomas Bloethe

Pag. 148

-------------------------------------------------------------------------------------------------ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY EDITED BY UMBERTO DESIDERI, GIAMPAOLO MANFRIDA, ENRICO SCIUBBA FIRENZE UNIVERSITY PRESS, 2012, ISBN ONLINE : 978-88-6655-322-9

» On the use of MPT to derive optimal RES electricity generation mixes …….... (ID 459) Paula Ferreira, Jorge Cunha

Pag. 164

» Stability and limit cycles in an exergy-based model of population …….... dynamics (ID 128) Enrico Sciubba, Federico Zullo

Pag. 179

» The influence of primary measures for reducing NOx emissions on …….... energy steam boiler efficiency (ID 125) Goran Stupar, Dragan Tucakovi , Titoslav Živanovi , Miloš Banjac, Sr an Beloševi ,Vladimir Beljanski, Ivan Tomanovi , Nenad Crnomarkovi , Miroslav Sijer

Pag. 193

» The Lethe city car of the University of Roma 1: final proposed …….... configuration (ID 45) Roberto Capata, Enrico Sciubba

Pag. 206

VIII. 2 POSTER SESSION » A variational optimization of a finite-time thermal cycle with a Stefan- …….... Boltzmann heat transfer law (ID 333) Juan C.Chimal-Eguia, Norma Sanchez-Salas

Pag. 216

» Modeling and simulation of a boiler unit for steam power plants (ID 545) Luca Moliterno, Claudia Toro

……....

Pag. 217

» Numerical Modelling of straw combustion in a moving bed combustor (ID 412) Biljana Miljkoviü, Ivan Pešenjanski, Borivoj Stepanov, Vladimir Milosavljeviü, Vladimir Rajs

……....

Pag. 218

» Physicochemical evaluation of the properties of the coke formed at …….... radiation area of light hydrocarbons pyrolysis furnace in petrochemical industry (ID 10) Jaqueline Saavedra Rueda , Angélica María Carreño Parra, María del Rosario Pérez Trejos, Dionisio Laverde Cataño, Diego Bonilla Duarte, Jorge Leonardo Rodríguez Jiménez, Laura María Díaz Burgos

Pag. 219

» Rotor TG cooled (ID 121) …….... Chiara Durastante, Paolo Petroni, Michela Spagnoli, Vincenzo Rizzica, Jörg Helge Wirfs

Pag. 220

» Study of the phase change in binary alloy (ID 534) Aroussia Jaouahdou, Mohamed J. Safi, Herve Muhr

……....

Pag. 221

» Technip initiatives in renewable energies and sustainable technologies (ID 527) Pierfrancesco Palazzo, Corrado Pigna

……....

Pag. 222

ix

----------------------------------------------------------------------CONTENTS OF ALL THE VOLUMES -----------------------------------------------------------------------

VOLUME I I . 1 - SIMULATION OF ENERGY CONVERSION SYSTEMS » A novel hybrid-fuel compressed air energy storage system for China’s situation (ID 531) Wenyi Liu, Yongping Yang, Weide Zhang, Gang Xu,and Ying Wu » A review of Stirling engine technologies applied to micro-cogeneration systems (ID 338) Ana C Ferreira, Manuel L Nunes, Luís B Martins, Senhorinha F Teixeira » An organic Rankine cycle off-design model for the search of the optimal control strategy (ID 295) Andrea Toffolo, Andrea Lazzaretto, Giovanni Manente, Marco Paci » Automated superstructure generation and optimization of distributed energy supply systems (ID 518) Philip Voll, Carsten Klaffke, Maike Hennen, André Bardow » Characterisation and classification of solid recovered fuels (SRF) and model development of a novel thermal utilization concept through air- gasification (ID 506) Panagiotis Vounatsos, Konstantinos Atsonios, Mihalis Agraniotis, Kyriakos D. Panopoulos, George Koufodimos,Panagiotis Grammelis, Emmanuel Kakaras » Design and modelling of a novel compact power cycle for low temperature heat sources (ID 177) Jorrit Wronski, Morten Juel Skovrup, Brian Elmegaard, Harald Nes Rislå, Fredrik Haglind » Dynamic simulation of combined cycles operating in transient conditions: an innovative approach to determine the steam drums life consumption (ID 439) Stefano Bracco » Effect of auxiliary electrical power consumptions on organic Rankine cycle system with low-temperature waste heat source (ID 235) Samer Maalouf, Elias Boulawz Ksayer, Denis Clodic » Energetic and exergetic analysis of waste heat recovery systems in the cement industry (ID 228) Sotirios Karellas, Aris Dimitrios Leontaritis, Georgios Panousis, Evangelos Bellos, Emmanuel Kakaras » Energy and exergy analysis of repowering options for Greek lignite-fired power plants (ID 230) Sotirios Karellas, Aggelos Doukelis, Grammatiki Zanni, Emmanuel Kakaras » Energy saving by a simple solar collector with reflective panels and boiler (ID 366) Anna Stoppato, Renzo Tosato » Exergetic analysis of biomass fired double-stage Organic Rankine Cycle (ORC) (ID 37) Markus Preißinger, Florian Heberle, Dieter Brüggemann » Experimental tests and modelization of a domestic-scale organic Rankine cycle (ID 156) Roberto Bracco, Stefano Clemente, Diego Micheli, Mauro Reini » Model of a small steam engine for renewable domestic CHP system (ID 31 ) Giampaolo Manfrida, Giovanni Ferrara, Alessandro Pescioni » Model of vacuum glass heat pipe solar collectors (ID 312) Daniele Fiaschi, Giampaolo Manfrida x

» Modelling and exergy analysis of a plasma furnace for aluminum melting process (ID 254) Luis Enrique Acevedo, Sergio Usón, Javier Uche, Patxi Rodríguez » Modelling and experimental validation of a solar cooling installation (ID 296) Guillaume Anies, Pascal Stouffs, Jean Castaing-Lasvignottes » The influence of operating parameters and occupancy rate of thermoelectric modules on the electricity generation (ID 314) Camille Favarel, Jean-Pierre Bédécarrats, Tarik Kousksou, Daniel Champier » Thermodynamic and heat transfer analysis of rice straw co-firing in a Brazilian pulverised coal boiler (ID 236) Raphael Miyake, Alvaro Restrepo, Fábio Kleveston Edson Bazzo, Marcelo Bzuneck » Thermophotovoltaic generation: A state of the art review (ID 88) Matteo Bosi, Claudio Ferrari, Francesco Melino, Michele Pinelli, Pier Ruggero Spina, Mauro Venturini I . 2 – HEAT AND MASS TRANSFER » A DNS method for particle motion to establish boundary conditions in coal gasifiers (ID 49) Efstathios E Michaelides, Zhigang Feng » Effective thermal conductivity with convection and radiation in packed bed (ID 60) Yusuke Asakuma » Experimental and CFD study of a single phase cone-shaped helical coiled heat exchanger: an empirical correlation (ID 375) Daniel Flórez-Orrego, Walter Arias, Diego López, Héctor Velásquez » Thermofluiddynamic model for control analysis of latent heat thermal storage system (ID 207) Adriano Sciacovelli, Vittorio Verda, Flavio Gagliardi » Towards the development of an efficient immersed particle heat exchanger: particle transfer from low to high pressure (ID 202) Luciano A. Catalano, Riccardo Amirante, Stefano Copertino, Paolo Tamburrano, Fabio De Bellis I . 3 – INDUSTRIAL ECOLOGY » Anthropogenic heat and exergy balance of the atmosphere (ID 122) Asfaw Beyene, David MacPhee, Ron Zevenhoven » Determination of environmental remediation cost of municipal waste in terms of extended exergy (ID 63) Candeniz Seckin, Ahmet R. Bayulken » Development of product category rules for the application of life cycle assessment to carbon capture and storage (537) Carlo Strazza, Adriana Del Borghi, Michela Gallo » Electricity production from renewable and non-renewable energy sources: a comparison of environmental, economic and social sustainability indicators with exergy losses throughout the supply chain (ID 247) Lydia Stougie, Hedzer van der Kooi, Rob Stikkelman » Exergy analysis of the industrial symbiosis model in Kalundborg (ID 218) Alicia Valero Delgado, Sergio Usón, Jorge Costa » Global gold mining: is technological learning overcoming the declining in ore grades? (ID 277) Adriana Domínguez, Alicia Valero

xi

» Personal transportation energy consumption (ID305) Matteo Muratori, Emmanuele Serra, Vincenzo Marano, Michael Moran » Resource use evaluation of Turkish transportation sector via the extended exergy accounting method (ID 43) Candeniz Seckin, Enrico Sciubba, Ahmet R. Bayulken » The impact of higher energy prices on socio-economic inequalities of German social groups (ID 80) Holger Schlör, Wolfgang Fischer, Jürgen-Friedrich Hake

VOLUME II II . 1 – EXERGY ANALYSIS AND 2ND LAW ANALYSIS » A comparative analysis of cryogenic recuperative heat exchangers based on exergy destruction (ID 129) Adina Teodora Gheorghian, Alexandru Dobrovicescu, Lavinia Grosu, Bogdan Popescu, Claudia Ionita » A critical exploration of the usefulness of rational efficiency as a performance parameter for heat exchangers (ID 307) Jim McGovern, Georgiana Tirca-Dragomirescu, Michel Feidt, Alexandru Dobrovicescu » A new procedure for the design of LNG processes by combining exergy and pinch analyses (ID 238) Danahe Marmolejo-Correa, Truls Gundersen » Advances in the distribution of environmental cost of water bodies through the exergy concept in the Ebro river (ID 258) Javier Uche Marcuello, Amaya Martínez Gracia, Beatriz Carrasquer Álvarez, Antonio Valero Capilla » Application of the entropy generation minimization method to a solar heat exchanger: a pseudo-optimization design process based on the analysis of the local entropy generation maps (ID 357) Giorgio Giangaspero, Enrico Sciubba » Comparative analysis of ammonia and carbon dioxide two-stage cycles for simultaneous cooling and heating (ID 84) Alexandru Dobrovicescu, Ciprian Filipoiu, Emilia Cerna Mladin, Valentin Apostol, Liviu Drughean » Comparison between traditional methodologies and advanced exergy analyses for evaluating efficiency and externalities of energy systems (ID 515) Gabriele Cassetti, Emanuela Colombo » Comparison of entropy generation figures using entropy maps and entropy transport equation for an air cooled gas turbine blade (ID 468) Omer Emre Orhan, Oguz Uzol » Conventional and advanced exergetic evaluation of a supercritical coal-fired power plant (ID 377) Ligang Wang, Yongping Yang, Tatiana Morosuk, George Tsatsaronis » Energy and exergy analyses of the charging process in encapsulted ice thermal energy storage (ID 164) David MacPhee, Ibrahim Dincer, Asfaw Beyene » Energy integration and cogeneration in nitrogen fertilizers industry: thermodynamic estimation of the efficiency, potentials, limitations and environmental impact. Part 1: energy integration in ammonia production plants (ID 303) Zornitza Vassileva Kirova-Yordanova » Evaluation of the oil and gas processing at a real production day on a North Sea oil platform using exergy analysis (ID 260) Mari Voldsund, Wei He, Audun Røsjorde, Ivar Ståle Ertesvåg, Signe Kjelstrup xii

» Exergetic and economic analysis of Kalina cycle for low temperature geothermal sources in Brazil (ID 345) Carlos Eymel Campos Rodriguez, José Carlos Escobar Palacios, Cesar Adolfo Rodríguez Sotomonte, Marcio Leme, Osvaldo José Venturini, Electo Eduardo Silva Lora, Vladimir Melián Cobasa, Daniel Marques dos Santos, Fábio R. Lofrano Dotto, Vernei Gialluca » Exergy analysis and comparison of CO2 heat pumps (ID 242) Argyro Papadaki, Athina Stegou - Sagia » Exergy analysis of a CO2 Recovery plant for a brewery (ID 72) Daniel Rønne Nielsen, Brian Elmegaard, C. Bang-Møller » Exergy analysis of the silicon production process (ID 118) Marit Takla, Leiv Kolbeinsen, Halvard Tveit, Signe Kjelstrup » Exergy based indicators for cardiopulmonary exercise test evaluation (ID 159) Carlos Eduardo Keutenedjian Mady, Cyro Albuquerque Neto, Tiago Lazzaretti Fernandes, Arnaldo Jose Hernandez, Paulo Hilário Nascimento Saldiva, Jurandir Itizo Yanagihara, Silvio de Oliveira Junior » Exergy disaggregation as an alternative for system disaggregation in thermoeconomics (ID 483) José Joaquim Conceição Soares Santos, Atilio Lourenço, Julio Mendes da Silva, João Donatelli, José Escobar Palacio » Exergy intensity of petroleum derived fuels (ID 117) Julio Augusto Mendes da Silva, Maurício Sugiyama, Claudio Rucker, Silvio de Oliveira Junior » Exergy-based sustainability evaluation of a wind power generation system (ID 542) Jin Yang, B. Chen, Enrico Sciubba » Human body exergy metabolism (ID 160) Carlos Eduardo Keutenedjian Mady, Silvio de Oliveira Junior » Integrating an ORC into a natural gas expansion plant supplied with a co-generation unit (ID 273) Sergio Usón, Wojciech Juliusz Kostowski » One-dimensional model of an optimal ejector and parametric study of ejector efficiency (ID 323) Ronan Killian McGovern, Kartik Bulusu, Mohammed Antar, John H. Lienhard » Optimization and design of pin-fin heat sinks based on minimum entropy generation (ID 6) Jose-Luis Zuniga-Cerroblanco, Abel Hernandez-Guerrero, Carlos A. Rubio-Jimenez, Cuauhtemoc Rubio-Arana, Sosimo E. Diaz-Mendez » Performance analysis of a district heating system (ID 271) Andrej Ljubenko, Alojz Poredoš, Tatiana Morosuk, George Tsatsaronis » System analysis of exergy losses in an integrated oxy-fuel combustion power plant (ID 64) Andrzej Zi bik, Pawe G adysz » What is the cost of losing irreversibly the mineral capital on Earth? (ID 220) Alicia Valero Delgado, Antonio Valero II . 2 – THERMODYNAMICS » A new polygeneration system for methanol and power based on coke oven gas and coal gas (ID 252) Hu Lin, Hongguang Jin, Lin Gao, Rumou Li » Argon-Water closed gas cycle (ID 67) Federico Fionelli, Giovanni Molinari » Binary alkane mixtures as fluids in Rankine cycles (ID 246) M. Aslam Siddiqi, Burak Atakan xiii

» Excess enthalpies of second generation biofuels (ID 308) Alejandro Moreau, José Juan Segovia, M. Carmen Martín, Miguel Ángel Villamañán, César R. Chamorro, Rosa M. Villamañán » Local stability analysis of a Curzon-Ahlborn engine considering the Van der Waals equation state in the maximum ecological regime (ID 281) Ricardo Richard Páez-Hernández, Pedro Portillo-Díaz, Delfino Ladino-Luna, Marco Antonio Barranco-Jiménez » Some remarks on the Carnot's theorem (ID 325) Julian Gonzalez Ayala, Fernando Angulo-Brown » The Dead State (ID 340) Richard A. Gaggioli » The magnetocaloric energy conversion (ID 97) Andrej Kitanovski, Jaka Tusek, Alojz Poredos

VOLUME III THERMO-ECONOMIC ANALYSIS AND OPTIMIZATION » A comparison of optimal operation of residential energy systems using clustered demand patterns based on Kullback-Leibler divergence (ID 142) Akira Yoshida, Yoshiharu Amano, Noboru Murata, Koichi Ito, Takumi Hashizume » A Model for Simulation and Optimal Design of a Solar Heating System with Seasonal Storage (ID 51) Gianfranco Rizzo » A thermodynamic and economic comparative analysis of combined gas-steam and gas turbine air bottoming cycle (ID 232) Tadeusz Chmielniak, Daniel Czaja, Sebastian Lepszy » Application of an alternative thermoeconomic approach to a two-stage vapor compression refrigeration cycle with intercooling (ID 135) Atilio Barbosa Lourenço, José Joaquim Conceição Soares Santos, João Luiz Marcon Donatelli » Comparative performance of advanced power cycles for low temperature heat sources (ID 109) Guillaume Becquin, Sebastian Freund » Comparison of nuclear steam power plant and conventional steam power plant through energy level and thermoeconomic analysis (ID 251) S. Khamis Abadi, Mohammad Hasan Khoshgoftar Manesh, M. Baghestani, H. Ghalami, Majid Amidpour » Economic and exergoeconomic analysis of micro GT and ORC cogeneration systems (ID 87) Audrius Bagdanavicius, Robert Sansom, Nick Jenkins, Goran Strbac » Exergoeconomic comparison of wet and dry cooling technologies for the Rankine cycle of a solar thermal power plant (ID 300) Philipp Habl, Ana M. Blanco-Marigorta, Berit Erlach » Influence of renewable generators on the thermo-economic multi-level optimization of a poly-generation smart grid (101) Massimo Rivarolo, Andrea Greco, Francesca Travi, Aristide F. Massardo » Local stability analysis of a thermoeconomic model of an irreversible heat engine working at different criteria of performance (ID 289) Marco A. Barranco-Jiménez, Norma Sánchez-Salas, Israel Reyes-Ramírez, Lev Guzmán-Vargas » Multicriteria optimization of a distributed trigeneration system in an industrial area (ID 154) Dario Buoro, Melchiorre Casisi, Alberto de Nardi, Piero Pinamonti, Mauro Reini xiv

» On the effect of eco-indicator selection on the conclusions obtained from an exergoenvironmental analysis (ID 275) Tatiana Morosuk, George Tsatsaronis, Christopher Koroneos » Optimisation of supply temperature and mass flow rate for a district heating network (ID 104) Marouf Pirouti, Audrius Bagdanavicius, Jianzhong Wu, Janaka Ekanayake » Optimization of energy supply systems in consideration of hierarchical relationship between design and operation (ID 389) Ryohei Yokoyama, Shuhei Ose » The fuel impact formula revisited (ID 279) Cesar Torres, Antonio Valero » The introduction of exergy analysis to the thermo-economic modelling and optimisation of a marine combined cycle system (ID 61) George G. Dimopoulos, Chariklia A. Georgopoulou, Nikolaos M.P. Kakalis » The relationship between costs and environmental impacts in power plants: an exergybased study (ID 272) Fontina Petrakopoulou, Yolanda Lara, Tatiana Morosuk, Alicia Boyano, George Tsatsaronis » Thermo-ecological evaluation of biomass integrated gasification gas turbine based cogeneration technology (ID 441) Wojciech Stanek, Lucyna Czarnowska, Jacek Kalina » Thermo-ecological optimization of a heat exchanger through empirical modeling (ID 501) Ireneusz Szczygie , Wojciech Stanek, Lucyna Czarnowska, Marek Rojczyk » Thermoeconomic analysis and optimization in a combined cycle power plant including a heat transformer for energy saving (ID 399) Elizabeth Cortés Rodríguez, José Luis Castilla Carrillo, Claudia A. Ruiz Mercado, Wilfrido Rivera Gómez-Franco » Thermoeconomic analysis and optimization of a hybrid solar-electric heating in a fluidized bed dryer (ID 400) Elizabeth Cortés Rodríguez, Felipe de Jesús Ojeda Cámara, Isaac Pilatowsky Figueroa » Thermoeconomic approach for the analysis of low temperature district heating systems (ID 208) Vittorio Verda, Albana Kona » Thermo-economic assessment of a micro CHP systems fuelled by geothermal and solar energy (ID 166) Duccio Tempesti, Daniele Fiaschi, Filippo Gabuzzini » Thermo-economic evaluation and optimization of the thermo-chemical conversion of biomass into methanol (ID 194) Emanuela Peduzzi, Laurence Tock, Guillaume Boissonnet, François Marechal » Thermoeconomic fuel impact approach for assessing resources savings in industrial symbiosis: application to Kalundborg Eco-industrial Park (ID 256) Sergio Usón, Antonio Valero, Alicia Valero, Jorge Costa » Thermoeconomics of a ground-based CAES plant for peak-load energy production system (ID 32) Simon Kemble, Giampaolo Manfrida, Adriano Milazzo, Francesco Buffa

xv

VOLUME IV IV . 1 - FLUID DYNAMICS AND POWER PLANT COMPONENTS » A control oriented simulation model of a multistage axial compressor (ID 444) Lorenzo Damiani, Giampaolo Crosa, Angela Trucco » A flexible and simple device for in-cylinder flow measurements: experimental and numerical validation (ID 181) Andrea Dai Zotti, Massimo Masi, Marco Antonello » CFD Simulation of Entropy Generation in Pipeline for Steam Transport in Real Industrial Plant (ID 543) Goran Vu kovi , Gradimir Ili , Mi a Vuki , Milan Bani , Gordana Stefanovi » Feasibility Study of Turbo expander Installation in City Gate Station (ID 168) Navid Zehtabiyan Rezaie, Majid Saffar-Avval » GTL and RME combustion analysis in a transparent CI engine by means of IR digital imaging (ID 460) Ezio Mancaruso, Luigi Sequino, Bianca Maria Vaglieco » Some aspects concerning fluid flow and turbulence modeling in 4-valve engines (ID 116) Zoran Stevan Jovanovic, Zoran Masonicic, Miroljub Tomic IV . 2 - SYSTEM OPERATION CONTROL DIAGNOSIS AND PROGNOSIS » Adapting the operation regimes of trigeneration systems to renewable energy systems integration (ID 188) Liviu Ruieneanu, Mihai Paul Mircea » Advanced electromagnetic sensors for sustainable monitoring of industrial processes (ID 145) Uroš Puc, Andreja Abina, Anton Jegli , Pavel Cevc, Aleksander Zidanšek » Assessment of stresses and residual life of plant components in view of life-time extension of power plants (ID 453) Anna Stoppato, Alberto Benato and Alberto Mirandola » Control strategy for minimizing the electric power consumption of hybrid ground source heat pump system (ID 244) Zoi Sagia, Constantinos Rakopoulos » Exergetic evaluation of heat pump booster configurations in a low temperature district heating network (ID 148) Torben Ommen, Brian Elmegaard » Exergoeconomic diagnosis: a thermo-characterization method by using irreversibility analysis (ID 523) Abraham Olivares-Arriaga, Alejandro Zaleta-Aguilar, Rangel-Hernández V. H, Juan Manuel Belman-Flores » Optimal structural design of residential cogeneration systems considering their operational restrictions (ID 224) Tetsuya Wakui, Ryohei Yokoyama » Performance estimation and optimal operation of a CO2 heat pump water heating system (ID 344) Ryohei Yokoyama, Ryosuke Kato, Tetsuya Wakui, Kazuhisa Takemura » Performances of a common-rail Diesel engine fuelled with rapeseed and waste cooking oils (ID 213) Alessandro Corsini, Valerio Giovannoni, Stefano Nardecchia, Franco Rispoli, Fabrizio Sciulli, Paolo Venturini xv i

» Reduced energy cost through the furnace pressure control in power plants (ID 367) Vojislav Filipovi , Novak Nedi , Saša Prodanovi » Short-term scheduling model for a wind-hydro-thermal electricity system (ID 464) Sérgio Pereira, Paula Ferreira, A. Ismael Freitas Vaz

VOLUME V V . 1 - RENEWABLE ENERGY CONVERSION SYSTEMS » A co-powered concentrated solar power Rankine cycle concept for small size combined heat and power (ID 276) Alessandro Corsini, Domenico Borello, Franco Rispoli, Eileen Tortora » A novel non-tracking solar collector for high temperature application (ID 466) Wattana Ratismith, Anusorn Inthongkhum » Absorption heat transformers (AHT) as a way to enhance low enthalpy geothermal resources (ID 311) Daniele Fiaschi, Duccio Tempesti, Giampaolo Manfrida, Daniele Di Rosa » Alternative feedstock for the biodiesel and energy production: the OVEST project (ID 98) Matteo Prussi, David Chiaramonti, Lucia Recchia, Francesco Martelli, Fabio Guidotti » Assessing repowering and update scenarios for wind energy converters (ID 158) Till Zimmermann » Biogas from mechanical pulping industry – potential improvement for increased biomass vehicle fuels (ID 54) Mimmi Magnusson, Per Alvfors » Biogas or electricity as vehicle fuels derived from food waste - the case of Stockholm (ID 27) Martina Wikström, Per Alvfors » Compressibility factor as evaluation parameter of expansion processes in organic Rankine cycles (ID 292) Giovanni Manente, Andrea Lazzaretto » Design of solar heating system for methane generation (ID 445) Lucía Mónica Gutiérrez, P. Quinto Diez, L. R. Tovar Gálvez » Economic feasibility of PV systems in hotels in Mexico (ID 346) Augusto Sanchez, Sergio Quezada » Effect of a back surface roughness on annual performance of an air-cooled PV module (ID 193) Riccardo Secchi, Duccio Tempesti, Jacek Smolka » Energy and exergy analysis of the first hybrid solar-gas power plant in Algeria (ID 176) Fouad Khaldi » Energy recovery from MSW treatment by gasification and melting technology (ID 393) Fabrizio Strobino, Alessandro Pini Prato, Diego Ventura, Marco Damonte » Ethanol production by enzymatic hydrolysis process from sugarcane biomass - the integration with the conventional process (ID 189) Reynaldo Palacios-Bereche, Adriano Ensinas, Marcelo Modesto, Silvia Azucena Nebra » Evaluation of gas in an industrial anaerobic digester by means of biochemical methane potential of organic municipal solid waste components (ID 57) Isabella Pecorini, Tommaso Olivieri, Donata Bacchi, Alessandro Paradisi, Lidia Lombardi, Andrea Corti, Ennio Carnevale

xv ii

» Exergy analysis and genetic algorithms for the optimization of flat-plate solar collectors (ID 423) Soteris A. Kalogirou » Experimental study of tar and particles content of the produced gas in a double stage downdraft gasifier (ID 487) Ana Lisbeth Galindo Noguera, Sandra Yamile Giraldo, Rene Lesme-Jaén, Vladimir Melian Cobas, Rubenildo Viera Andrade, Electo Silva Lora » Feasibility study to realize an anaerobic digester fed with vegetables matrices in central Italy (ID 425) Umberto Desideri, Francesco Zepparelli, Livia Arcioni, Ornella Calderini, Francesco Panara, Matteo Todini » Investigations on the use of biogas for small scale decentralized CHP applications with a focus on stability and emissions (ID 140) Steven MacLean, Eren Tali, Anne Giese, Jörg Leicher » Kinetic energy recovery system for sailing yachts (ID 427) Giuseppe Leo Guizzi, Michele Manno » Mirrors in the sky: status and some supporting materials experiments (ID 184) Noam Lior » Numerical parametric study for different cold storage designs and strategies of a solar driven thermoacoustic cooler system (ID 284) Maxime Perier-Muzet, Pascal Stouffs, Jean-Pierre Bedecarrats, Jean Castaing-Lasvignottes » Parabolic trough photovoltaic/thermal collectors. Part I: design and simulation model (ID 102) Francesco Calise, Laura Vanoli » Parabolic trough photovoltaic/thermal collectors. Part II: dynamic simulation of a solar trigeneration system (ID 488) Francesco Calise, Laura Vanoli » Performance analysis of downdraft gasifier - reciprocating engine biomass fired smallscale cogeneration system (ID 368) Jacek Kalina » Proposing offshore photovoltaic (PV) technology to the energy mix of the Maltese islands (ID 262) Kim Trapani, Dean Lee Millar » Research of integrated biomass gasification system with a piston engine (ID 414) Janusz Kotowicz, Aleksander Sobolewski, Tomasz Iluk » Start up of a pre-industrial scale solid state anaerobic digestion cell for the co-treatment of animal and agricultural residues (ID 34) Francesco Di Maria, Giovanni Gigliotti, Alessio Sordi, Caterina Micale, Luisa Massaccesi » The role of biomass in the renewable energy system (ID 390) Ruben Laleman, Ludovico Balduccio, Johan Albrecht » Vegetable oils of soybean, sunflower and tung as alternative fuels for compression ignition engines (ID 500) Ricardo Morel Hartmann, Nury Nieto Garzón, Eduardo Morel Hartmann, Amir Antonio Martins Oliveira Jr, Edson Bazzo, Bruno Okuda, Joselia Piluski » Wind energy conversion performance and atmosphere stability (ID 283) Francesco Castellani, Emanuele Piccioni, Lorenzo Biondi, Marcello Marconi V. 2 - FUEL CELLS » Comparison study on different SOFC hybrid systems with zero-CO2 emission (ID 196) Liqiang Duan, Kexin Huang, Xiaoyuan Zhang and Yongping Yang xv iii

» Exergy analysis and optimisation of a steam methane pre-reforming system (ID 62) George G. Dimopoulos, Iason C. Stefanatos, Nikolaos M.P. Kakalis » Modelling of a CHP SOFC power system fed with biogas from anaerobic digestion of municipal wastes integrated with a solar collector and storage units (ID 491) Domenico Borello, Sara Evangelisti, Eileen Tortora

VOLUME VI VI . 1 - CARBON CAPTURE AND SEQUESTRATION » A novel coal-based polygeneration system cogenerating power, natural gas and liquid fuel with CO2 capture (ID 96) Sheng Li, Hongguang Jin, Lin Gao » Analysis and optimization of CO2 capture in a China’s existing coal-fired power plant (ID 532) Gang Xu, Yongping Yang, Shoucheng Li, Wenyi Liu and Ying Wu » Analysys of four-end high temperature membrane air separator in a supercritical power plant with oxy-type pulverized fuel boiler (ID 442) Janusz Kotowicz, Sebastian Stanis aw Michalski » Analysis of potential improvements to the lignite-fired oxy-fuel power unit (ID 413) Marcin Liszka, Jakub Tuka, Grzegorz Nowak, Grzegorz Szapajko » Biogas Upgrading: Global Warming Potential of Conventional and Innovative Technologies (ID 240) Katherine Starr, Xavier Gabarrell Durany, Gara Villalba Mendez, Laura Talens Peiro, Lidia Lombardi » Capture of carbon dioxide using gas hydrate technology (ID 103) Beatrice Castellani, Mirko Filipponi, Sara Rinaldi, Federico Rossi » Carbon dioxide mineralisation and integration with flue gas desulphurisation applied to a modern coal-fired power plant (ID 179) Ron Zevenhoven, Johan Fagerlund, Thomas Björklöf, Magdalena Mäkelä, Olav Eklund » Carbon dioxide storage by mineralisation applied to a lime kiln (ID 226) Inês Sofia Soares Romão, Matias Eriksson, Experience Nduagu, Johan Fagerlund, Licínio Manuel Gando-Ferreira, Ron Zevenhoven » Comparison of IGCC and CFB cogeneration plants equipped with CO2 removal (ID 380) Marcin Liszka, Tomasz Malik, Micha Budnik, Andrzej Zi bik » Concept of a “capture ready” combined heat and power plant (ID 231) Piotr Henryk Lukowicz, Lukasz Bartela » Cryogenic method for H2 and CH4 recovery from a rich CO2 stream in pre-combustion CCS schemes (ID 508) Konstantinos Atsonios, Kyriakos D. Panopoulos, Angelos Doukelis, Antonis Koumanakos, Emmanuel Kakaras » Design and optimization of ITM oxy-combustion power plant (ID 495) Surekha Gunasekaran, Nicholas David Mancini, Alexander Mitsos » Implementation of a CCS technology: the ZECOMIX experimental platform (ID 222) Antonio Calabrò, Stefano Cassani, Leandro Pagliari, Stefano Stendardo » Influence of regeneration condition on cyclic CO2 capture using pre-treated dispersed CaO as high temperature sorbent (ID 221) Stefano Stendardo, Antonio Calabrò

xix

» Investigation of an innovative process for biogas up-grading – pilot plant preliminary results (ID 56) Lidia Lombardi, Renato Baciocchi, Ennio Antonio Carnevale, Andrea Corti, Giulia Costa, Tommaso Olivieri, Alessandro Paradisi, Daniela Zingaretti » Method of increasing the efficiency of a supercritical lignite-fired oxy-type fluidized bed boiler and high-temperature three - end membrane for air separation (ID 438) Janusz Kotowicz, Adrian Balicki » Monitoring of carbon dioxide uptake in accelerated carbonation processes applied to air pollution control residues (ID 539) Felice Alfieri, Peter J Gunning, Michela Gallo, Adriana Del Borghi, Colin D Hills » Process efficiency and optimization of precipitated calcium carbonate (PCC) production from steel converter slag (ID 114) Hannu-Petteri Mattila, Inga Grigali nait , Arshe Said, Sami Filppula, Carl-Johan Fogelholm, Ron Zevenhoven » Production of Mg(OH)2 for CO2 Emissions Removal Applications: Parametric and Process Evaluation (ID 245) Experience Ikechukwu Nduagu, Inês Romão, Ron Zevenhoven » Thermodynamic analysis of a supercritical power plant with oxy type pulverized fuel boiler, carbon dioxide capture system (CC) and four-end high temperature membrane air seprator (ID 411) Janusz Kotowicz, Sebastian Stanis aw Michalski VI . 2 - PROCESS INTEGRATION AND HEAT EXCHANGER NETWORKS » A multi-objective optimization technique for co- processing in the cement production (ID 42) Maria Luiza Grillo Renó, Rogério José da Silva, Mirian de Lourdes Noronha Motta Melo, José Joaquim Conceição Soares Santos » Comparison of options for debottlenecking the recovery boiler at kraft pulp mills – Economic performance and CO2 emissions (ID 449) Johanna Jönsson, Karin Pettersson, Simon Harvey, Thore Berntsson » Demonstrating an integral approach for industrial energy saving (ID 541) René Cornelissen, Geert van Rens, Jos Sentjens, Henk Akse, Ton Backx, Arjan van der Weiden, Jo Vandenbroucke » Maximising the use of renewables with variable availability (ID 494) Andreja Nemet, Jiri Jaromír Klemeš, Petar Sabev Varbanov, Zdravko Kravanja » Methodology for the improvement of large district heating networks (ID 46) Anna Volkova, Vladislav Mashatin, Aleksander Hlebnikov, Andres Siirde » Optimal mine site energy supply (ID 306) Monica Carvalho, Dean Lee Millar » Simulation of synthesis gas production from steam oxygen gasification of Colombian bituminous coal using Aspen Plus® (ID 395) John Jairo Ortiz, Juan Camilo González, Jorge Enrique Preciado, Rocío Sierra, Gerardo Gordillo

VOLUME VII VII . 1 - BUILDING, URBAN AND COMPLEX ENERGY SYSTEMS » A linear programming model for the optimal assessment of sustainable energy action plans (ID 398) Gianfranco Rizzo, Giancarlo Savino xx

» A natural gas fuelled 10 kW electric power unit based on a Diesel automotive internal combustion engine and suitable for cogeneration (ID 477) Pietro Capaldi » Adjustment of envelopes characteristics to climatic conditions for saving heating and cooling energy in buildings (ID 430) Christos Tzivanidis, Kimon Antonopoulos, Foteini Gioti » An exergy based method for the optimal integration of a building and its heating plant. Part 1: comparison of domestic heating systems based on renewable sources (ID 81) Marta Cianfrini, Enrico Sciubba, Claudia Toro » Analysis of different typologies of natural insulation materials with economic and performances evaluation of the same buildings (ID 28) Umberto Desideri, Daniela Leonardi, Livia Arcioni » Complex networks approach to the Italian photovoltaic energy distribution system (ID 470) Luca Valori, Giovanni Luca Giannuzzi, Tiziano Squartini, Diego Garlaschelli, Riccardo Basosi » Design of a multi-purpose building "to zero energy consumption" according to European Directive 2010/31/CE: Architectural and plant solutions (ID 29) Umberto Desideri, Livia Arcioni, Daniela Leonardi, Luca Cesaretti ,Perla Perugini, Elena Agabitini, Nicola Evangelisti » Effect of initial systems on the renewal planning of energy supply systems for a hospital (ID 107) Shu Yoshida, Koichi Ito, Yoshiharu Amano, Shintaro Ishikawa, Takahiro Sushi, Takumi Hashizume » Effects of insulation and phase change materials (PCM) combinations on the energy consumption for buildings indoor thermal comfort (ID 387) Christos Tzivanidis, Kimon Antonopoulos, Eleutherios Kravvaritis » Energetic evaluation of a smart controlled greenhouse for tomato cultivation (ID 150) Nickey Van den Bulck, Mathias Coomans, Lieve Wittemans, Kris Goen, Jochen Hanssens, Kathy Steppe, Herman Marien, Johan Desmedt » Energy networks in sustainable cities: temperature and energy consumption monitoring in urban area (ID 190) Luca Giaccone, Alessandra Guerrisi, Paolo Lazzeroni and Michele Tartaglia » Extended exergy analysis of the economy of Nova Scotia, Canada David C Bligh, V.Ismet Ugursal

(ID 215)

» Feasibility study and design of a low-energy residential unit in Sagarmatha Park (Nepal) for envirnomental impact reduction of high altitude buildings (ID 223) Umberto Desideri, Stefania Proietti, Paolo Sdringola, Elisa Vuillermoz » Fire and smoke spread in low-income housing in Mexico (ID 379) Raul R. Flores-Rodriguez, Abel Hernandez-Guerrero, Cuauhtemoc Rubio-Arana, Consuelo A. Caldera-Briseño » Optimal lighting control strategies in supermarkets for energy efficiency applications via digital dimmable technology (ID 136) Salvador Acha, Nilay Shah, Jon Ashford, David Penfold » Optimising the arrangement of finance towards large scale refurbishment of housing stock using mathematical programming and optimisationg (ID 127) Mark Gerard Jennings, Nilay Shah, David Fisk » Optimization of thermal insulation to save energy in buildings (ID 174) Milorad Boji , Marko Mileti , Vesna Marjanovi , Danijela Nikoli , Jasmina Skerli » Residential solar-based seasonal thermal storage system in cold climate: building envelope and thermal storage (ID 342) Alexandre Hugo and Radu Zmeureanu

xxi

» Simultaneous production of domestic hot water and space cooling with a heat pump in a Swedish Passive House (ID 55) Johannes Persson, Mats Westermark » SOFC micro-CHP integration in residential buildings (ID 201) Umberto Desideri, Giovanni Cinti, Gabriele Discepoli, Elena Sisani, Daniele Penchini » The effect of shading of building integrated photovoltaics on roof surface temperature and heat transfer in buildings (ID 83) Eftychios Vardoulakis, Dimitrios Karamanis » The influence of glazing systems on energy performance and thermal comfort in nonresidential buildings (ID 206) Cinzia Buratti, Elisa Moretti, Elisa Belloni » Thermal analysis of a greenhouse heated by solar energy and seasonal thermal energy storage in soil (ID 405) Yong Li, Jin Xu, Ru-Zhu Wang » Thermodynamic analysis of a combined cooling, heating and power system under part load condition (ID 476) Qiang Chen, Jianjiao Zheng, Wei Han, Jun Sui, Hong-guang Jin VII . 2 - COMBUSTION, CHEMICAL REACTORS » Baffle as a cost-effective design improvement for volatile combustion rate increase in biomass boilers of simple construction (ID 233) Borivoj Stepanov, Ivan Pešenjanski, Biljana Miljkovi » Characterization of CH4-H2-air mixtures in the high-pressure DHARMA reactor (ID 287) Vincenzo Moccia, Jacopo D'Alessio » Development of a concept for efficiency improvement and decreased NOx production for natural gas-fired glass melting furnaces by switching to a propane exhaust gas fired process (ID 146) Jörn Benthin, Anne Giese » Experimental analysis of inhibition phenomenon management for Solid Anaerobic Digestion Batch process (ID 348) Francesco Di Maria, Giovanni Gigliotti, Alessio Sordi, Caterina Micale, Claudia Zadra, Luisa Massaccesi » Experimental investigations of the combustion process of n-butanol/diesel blend in an optical high swirl CI engine (ID 85) Simona Silvia Merola, G. Valentino, C. Tornatore, L. Marchitto , F. E. Corcione » Flameless oxidation as a means to reduce NOx emissions in glass melting furnaces (ID 141) Jörg Leicher, Anne Giese » Mechanism of damage by high temperature of the tubes, exposed to the atmosphere characteristic of a furnace of pyrolysis of ethane for ethylene production in the petrochemical industry (ID 65) Jaqueline Saavedra Rueda, Francisco Javier Perez Trujillo, Lourdes Isabel Meriño Stand, Harbey Alexi Escobar, Luis Eduardo Navas, Juan Carlos Amezquita » Steam reforming of methane over Pt/Rh based wire mesh catalyst in single channel reformer for small scale syngas production (ID 317) Haftor Orn Sigurdsson, Søren Knudsen Kær

xxii

ECOS 2012

VOLUME VIII

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

A Multi-Criteria Decision Analysis Tool to Support Electricity Planning ´ c Fernando Ribeiro a , Paula Ferreira b , Madalena Araujo a Department of Production and Systems, University of Minho, Guimar˜aes, Portugal,

[email protected] b Department of Production and Systems, University of Minho, Guimar˜aes, Portugal, [email protected] c Department of Production and Systems, University of Minho, Guimar˜aes, Portugal,

[email protected] Abstract: A Multi-Criteria Decision Analysis (MCDA) tool was designed to support the evaluation of different electricity production scenarios. The MCDA tool is implemented in Excel worksheet and uses information obtained from a mixed integer optimization model. Given the input, the MCDA allowed ranking different scenarios relying on their performance on 13 criteria covering economic, job market, quality of life of local populations, technical and environmental issues. The criteria were weighted using both direct weights and trade-off analysis. In this paper, scenarios for the case of the Portuguese electricity system are presented, as well as the results of the evaluation, using the MCDA tool, relying on the input from a group of academics with background in economics, engineering and environment.

Keywords: Energy decision making, electricity generation, MCDA, Sustainable Development.

1. Introduction Over the last two decades, international treaties, such as Kyoto Protocol, have been signed, and strategies to mitigate CO2 emissions have arisen in all the developed world nations. At the same time, Sustainable Development is becoming part of political discourse in the European Union. According to the European Union Sustainable Development Strategy (EUSDS), Sustainable Development envisages the ”continuous improvement of the quality of life of citizens through sustainable communities that manage and use resources efficiently and tap the ecological and social innovation potential of the economy, so as to ensure prosperity, environmental protection and social cohesion” [1]. As a result, the electricity production planning gets more constrained than before, resulting in a multi-objective problem [2]. What traditionally was simply a cost minimizing problem should now be evaluated also under Sustainable Development criteria. In this paper a Multi-Criteria Decision Analysis tool, designed for the evaluation of different electricity generation scenarios, is presented. When using multi-criteria decision methodologies, one has to have in mind that best solutions for some decision makers may not be universal best solutions, as results are made upon personal judgement of different criteria. In the present work, a panel of experts on energy systems was invited to map the diversity of opinions and preferences for the future of the Portuguese electricity system. The use of the MCDA tool was demonstrated for the evaluation of possible electricity scenarios drawn for Portugal in 2020. The criteria used cover Sustainable Development (social, cost and environmental) issues among others like visual impacts and technical issues of power systems, as addressed in section 3.2. The criteria were drawn from both interviews conducted in previous work [3] and from the literature. Figure 1 summarizes the methodological approach to the problem. The two main blocks of the methodology are Scenario Generation and Scenario Evaluation (MCDA Tool). Sections 2 and 3 Corresponding author: Paula Ferreira, Email: [email protected]

1

Figure 1: Evaluation of scenarios for electricity production, with MCDA evaluation are dedicated to each one of these topics. As the Scenario Generation addesses the future of the Portuguese power generation system, the remainder of this section overviews this particular case.

1.1. Power Generation in Portugal Electricity in Portugal is mainly generated from large hydro, thermal and wind power, as can be seen in Figure 2. Thermal power is mostly provided with coal and CCGT (combined cycle gas turbines) power plants. Special Regime Production include all the technologies benefiting from feed-in tariffs, which are in Figure 2 divided in Wind power and ”Other SRP”. The Portuguese electricity system is strongly influenced by the rainfall characteristics. Although the large hydro power installed capacity remained almost unchanged between 2006 and 2010, in fact the hydro electricity production suffered strong variations.1 In 2007, the Portuguese state launched a new plan for installing more hydro power, known as PNBEPH (Plano Nacional de Barragens de Elevado Potencial Hidroel´ectrico)[4]. It aimed to reduce the unused hydro power potential from 54% to 33% until 2020, installing new 2059 MW. This was expected to be achieved by two means: increasing installed power of already existing facilities (909 MW), and building ten new hydro power plants totaling 1150 MW of installed power. Among these projects, some include pumping capacity. The use of pumping was justified to the need to complement additional wind power to be installed: given that wind farms may produce more in off-peak hours when electricity prices are lower, this energy can be used to pump water back to dams, so that hydro power can be generated during the hours of higher consumption and higher electricity prices. In 2007 the PNBEPH forecasted that in 2010 there would be 5100 MW of installed wind power, which contrasted 1 The yearly variation of hydro power production is reflected on the so-called ”hydraulicity factor”, which for an average year the equals 1.

2

Figure 2: Installed power in Portugal, 2010. Own elaboration from www.ren.pt data. ”Other SRP” include non-renewable and renewable cogeneration, biomass, small hydro, photovoltaics and wave power. with the 3751 MW achieved in reality [5]. As a result, the completion of these plans is constrained by political and other factors (such as the fall of electricity consumption in 2010 and 2011). The future of the Portuguese power system remains uncertain, and in section 2.3 some possible scenarios for 2020 are explored.

Figure 3: Electricity generation in Portugal, 2010. Own elaboration from www.ren.pt data. In order to present the numbers for a typical rainfall year, the numbers for hydro power were divided by the hidraulicity factor, which in 2010 was 1.31 [6]. The exceeding energy was assumed to be covered equally by coal and natural gas.

2. Scenario Generation 2.1. Model description In this section the Scenario Generation phase of the methodology mentioned in Figure 1 is addressed. In short, a Mixed Integer Linear Programming (MILP) model, programmed in GAMS (General Algebraic Modeling System) was used. The input data is given in an Excel file, as well as the final results.

3

For the detailed description of the used model, see [7]. The source code was used to create scenarios with different characteristics, based on the cost optimization of the electricity system. These scenarios represent different possible futures for the Portuguese power generation system in a 10 year range, departing from the present characteristics of the system. A scenario is charaterized by a set of newly installed power plants of each technology, that, together with the already installed ones, will supply the electricity demand. The technologies considered as variables were hydro power, wind, natural gas and coal; on the other hand non-wind Special Regime Production was assumed to remain constant for every scenario. The remainder of this subsection contains complementary information of the given reference [7]. The demand and peak load data are presented in the Excel input file. The scenarios depend on the demand of electricity, Dt,m , which were computed according to recent forecasts, information available in the Portuguese National Renewable Energy Action Plans [8]. According to this data, demand, which was about 52 TWh in 2010, will increase 12 TWh in 10 years. The rate of the peak load growth was adjusted accordingly to the rate of consumption growth. The present values of non-Wind Special Regime Production (SRP) installed power and generated energy, as well as expected growth are computed in the excel input worksheet, according to the information collected in the report available in the Portuguese Renewable Action Plan ([8], pages 117 and 118). Non-wind SRP includes the following technologies: non-renewable cogeneration, biomass, small hydro, photovoltaics and wave power. Therefore, a new parameter was added in the code, srp renewable ratiot,m , to express the monthly percentage of renewable energy among the SRP. As addressed later in this section, this value is necessary to calculate the percentage of renewable energy generated in a given solution: srp renewable ratiot,m,i = 1 −

PS RPt,m,i=non renewable cogeneration PS RPt,m,i

(1)

where PS RPt,m,i refers to the energy generated from SRP source i, in the month m of the year t. In order to account for the CO2 emissions of SRP, the monthly generation of non-renewable cogeneration was multiplied by the same CO2 emissions factor that affects CCGT groups. The value of srp average emissions was thus calculated in order to express the emissions from the SRP in the planning period (2011 to 2020). For calculating the SRP costs, [9] values were used (exchange ratio of 1 USD = 0.7325 EURO). From these values, the overall SRP levelized costs, srp levelized cost were obtained, for the whole planning period: X Pt,m,i (2) ci srp levelized cost = PS RPt,m,i t,m,i where ci stands for the levelized cost for each SRP technology and Pt,m,i is the monthly energy produced by SRP technology i in the month m of year t.

2.2. Scenarios A variety of scenarios to use in the MCDA tool can be generated, and these are solutions for the model. In table 1, five possible scenarios of electricity generation in the year 2020 are presented, aiming to represent five different strategies, representative of different energy policy trends: investment in natural gas, investment in coal, investment in a mix of hydro and gas, investment in a mix of hydro and wind, and a moderated scenario following a business-as-usual approach. Obviously, none of these scenarios is likely to happen in this exact form due to the infinity of possible and distinct

4

combinations. However, given the present state of the Portuguese electricity system, these are five possible strategies representative of different energy policy trends. The evaluation of more scenarios demands additional input information and higher response time on the MCDA tool. In order to ensure the effective participation of experts it was decided to keep the number of scenarios low. As the objective function of the model is the minimization of the costs, different constraints used to diversify the scenarios were created. These constraints were of two types: allowing the program to install or not power plants of a specific technology, and, on the other hand, a renewable energy quota to be met in 2020. Not using these constraints would result in the model covering the growing demand by installing only new coal power plants, the least costly solution. Table 1: Characterization of scenarios

Scenario

Base

Constraints Minimum New inRenewable stalled Quota technologies 45% All technologies allowed

Natural Gas

Turned off

Coal

Turned off

Hydro-Gas

45%

Maximum Renewable

70%

Only CCGT allowed Turned off Only CCGT and hydro power allowed No coal or CCGT allowed

New installed power

700MW coal, 1000MW hydro, 4400MW wind, 1180MW other SRP (all SRP excluding wind power) 2350MW natural gas, 1180MW other SRP

Results Cost Emissions (euro (CO2 ton per per GWh) MWh) 25.69 262

External energy Dependency 30%

25.24

294

53%

2550MW coal, 1180MW other SRP 2050MW natural gas, 2000MW hydro, 1180MW other SRP

23.75

360

55%

25.96

286

45%

2000MW hydro, 4400MW wind, 1180MW other SRP

26.37

250

28%

The ”Coal” scenario is the least costly one, but also leads to the highest external energy dependency (that is, highest share of coal and natural gas) and presents the highest CO2 emissions. The other extreme case, presenting lowest external energy dependency and less CO2 emissions is the ”Maximum Renewable” scenario, which costs are about 11% higher than for the ”Coal” scenario.

3. Scenario Evaluation Using the Multi-Criteria Decision Analysis Tool The MCDA tool2 is presented on an Excel worksheet and aims to rank the suitability of electricity production scenarios according to 13 criteria. In the remainder of this section, firstly the methodology is exposed, then the MCDA tool is presented and finally applied to a case study, using the five scenarios presented in the previous section.

3.1. Methodology A vast literature for MCDA applications to energy planning exists (see for example [10] and [11] for an overview). The proposed methodology could be summarized as direct weighting with an additive 2

The tool is available for download in http://sepp.dps.uminho.pt/.

5

value function for amalgamation. As a result, it involves three phases, already mentioned in Figure 1: Impact Evaluation, Direct Weighting and Trade-off Analysis. Impact Evaluation is the phase where a score, score s,c is assigned to each scenario s and criteria c. These values are then normalized, using a linear function v s,c , so that the best values become 1 and the worst values become 0. The user then assigns directly weights wc to each criteria c. Finally, for every criteria c, trade-offs are presented in terms of costs, while the user is still able to change weights according to his perceptions. The final value for the scenario s is calculated according to the Additive Value Function (AVF), as follows: AV F s =

X

wci × v s,ci

(3)

where the higher the value, the better the solution is. A brief example is now presented to illustrate the calculation of a trade-off: consider, from the above scenarios, that the user is weighting only two criteria: costs and external dependency. Taking into account that ”Coal” presents least cost and highest energy dependency, the opposite case of ”Maximum Renewable”, the normalization of these criteria would consist in vcoal,cost =1, vmax renew,cost =0, vcoal,dependency =0, vmax renew,dependency =1. As can be seen in Table 2, if only two criteria are weighted and the user gives the same importance to the costs and the energy dependency, he assumes implicitly that for him it is indifferent to choose scenario ”Coal” or ”Maximum Renewable” scenarios. Here the notion of trade-off appears: for the user, the energy dependency of the ”Maximum Renewable” scenario is worth 2,62 euro/MWh, which is the difference in cost between the scenario ”Maximum Renewable” and ”Coal” (26,37 minus 23,75). The calculation of the trade-off T s,c is performed according to the following equation: T s,c =

wc × score s,c × (26, 37 − 23, 75) wcost

(4)

Since T s,c is already multiplied by the range of the price (the parcel on the right), its value is given in euro/MWh. The user is always given the % of the costs that this increment represents in relation of the coal solution cost: in the case of the example where costs and dependency have the same weight, T=2,62 euro/MWh and 2,62/23,75 equals 11,01%. It is worthy observing that when the weight of the cost is equal to the weight of the external energy dependency, the scenario with best performance is the ”Base”, with AVF=94,79. In case the user gives the costs a weight twice the energy dependency, he would value the energy dependency in 1,31 euro/MWh (or 5,5%) and in this case the ”Coal” scenario performs better than any other.

6

Table 2: Calculation of additive value function (AVF) by weighting two criteria

Coal 23,75 1 55% 0

Scenario s Hydro-Gas 25,96 0,15 47% 0,3

Criteria c score s,cost v s,cost score s,dependency v s,dependency

Base 25,69 0,26 30% 0,93

Natural Gas 25,24 0,43 53% 0,07

AV F s

94,79

40,47

wcost =wdependency =80 80 36,09

80 50 80

AV F s

72,19

46,88

wcost =100, wdependency =50 100 30,30

AV F s

84,43

23,20

wcost =40, wdependency =80 40 29,90

Maximum Renewable 26,37 0 28% 1

3.2. The MCDA tool The proposed MCDA tool is presented in an Excel Workbook with five Sheets, as follows: 1. General Instructions The purpose of the tool is presented, as well as a summary of each of the following pages. 2. Scenarios The scenarios are presented in the form of graphics of installed power and produced electricity. Energy dependency ratio, CO2 emissions and annualized costs are also displayed graphically. 3. Instructions Instructions for the following sheet are presented, along with an example. 4. Impact Evaluation and Weighting Here the user is presented with the 13 criteria, along with explanations of every one of them. The user then fills the required cells, according to what he percepts to be the impacts generated by each scenario. Trade-offs are presented. 5. Results Results are printed: both ranking of scenarios and contribution of each criterion is given. In the remainder of this section the information on the sheet Impact Evaluation and Weighting is introduced. The criteria, Ci , and their description, are given as follows in Table 3. Since not all the impacts can be easily agreed upon, it was decided that the user might play a role on valuing them, as detailed in Table 3, column ”Scenario score i s,c ”. Information of investment, operation & maintenance of the whole group of power plants is included in a single cost criterion. Positive impacts in industry, job creation and dependency on foreign fossil fuels have been an international concern for sustainable energy decisions [11] [10] with implications at national level [8]. Diversification of the electricity mix is also seen as important for sustainability goals [12] contributing to the security of supply. Local income, visual and noise impacts, as well as land use and public health were identified as important issues for local populations’ standards of living, by the authors [13]. It is sometimes argued that the intermittency of the renewables imply they are overrated in levelized costs [14]: therefore, a criteria which accounts for the dispatchable rate of power on each solution was included. According to [15], the transmission system expansion requirements may be larger when renewable energy shares are higher; as the scenarios vary respecting

7

to that aspect, the criteria was proposed to be evaluated. Given the importance that CO2 emissions play in the economy nowadays, this criterion was also included.

Table 3: Description of the criteria used in the MCDA Ci C1

Name Costs

C2

National Industry

C3

Energy dency

C4

Employment

C5

Visual Impact

Impact caused by the construction of new power plants upon the sightseeing.

C6

Noise

Noise impact caused in neighbor areas by the new infra-structures.

Depen-

Description Sum of fixed and variable costs, divided by the total electricity produced during the planning period. The fixed costs are related with the investment cost applied to the new power plants and also with all fixed O&M costs. The variable costs include fuel and variable O&M costs for new and previously installed power plants. Impact of the scenario on the dynamics of the national industry. Rate of dependency on foreign sources in year 2020, calculated as the sum of energy produced in thermal power plants (coal, natural gas and non-renewable cogeneration) divided by the total energy amount produced. Employment created by the construction, operation and maintenance of the power plants.

8

Scenario score i s,c Values in e/MWh, obtained from the MILP model. User can not change values.

Score in ordinal scale, ranging from 1 (worst) to 5 (best). Requires user to attribute values according to own perception. Values in %, obtained from the MILP model. User can not change values.

Values are number of jobs. Obtained from the MILP model, based on [16]. Although values are given, the user may attribute different values according to own perception. Score in ordinal scale, ranging from 1 (worst) to 5 (best). Requires user to attribute values according to own perception. Score in ordinal scale, ranging from 1 (worst) to 5 (best), based on [17]. Although values are given, user may attribute values different according to own perception.

C7

Local Income

Rents originated by land use, for both public and private sectors.

C8

Diversity of Mix

C9

Rate of Dispatchable Power

C10

Investment Transmission Network

C11

CO2 Emissions

Diversity of installed power, calculated according to the Shannon-Wiener Index. Ratio between the sum of installed power of coal, CCGT, dam hydro power plants, and all the installed power. Additional investments required by the scenario. It was assumed that wind power has the worst impact, followed by hydro power, and no additional investment is required by natural gas and coal power plants. Ratio between CO2 emissions and the total electricity generated in the overall planning period.

C12

Land Use

C13

Public Health

in

Amount of land which becomes unusable by the scenario. Contamination of air, water, and general impact on public health.

Score in ordinal scale, ranging from 1 (worst) to 5 (best). Requires user to attribute values according to own perception. Higher values are better. Obtained from the MILP model, based on [18]. User can not change values. Score is given in %. Obtained from the MILP model. User does not change values. Score in ordinal scale, ranging from 1 (worst) to 5 (best). Although the values are given, the user may attribute different values according to own perception.

Values are given in tons of CO2 per GWh of electricity produced in the planning period. Obtained from the MILP model. User can not change values. Values are given in 1000 km2, based on [16]. Obtained from the MILP model. User can not change values. Score is based on [17]. Obtained from the MILP model. User can not change values.

Figure 4 presents an example of the user’s views of the MCDA tool for the C2 criterion (National Industry). The scale for this criterion ranges from 1 (Low dynamics in industry) to 5 (Leadership of industry, resulting in capacity for exporting), and the user has assigned the following impacts for I s,c : Ibase,national industry =4, Inatural gas,national industry =2, Icoal,national industry =2, Ihydro−gas,national industry =3, Imaximum renewable,national industry =5. The blue cell is the weight of the criterion, assigned as 20 in the example. The information displayed in the plot indicates that the user accepts to increase the costs in 2.20%, in order to increase the national industry dynamics from score 2 to score 5. In other words, the user wishes to increase dynamics national industry from ”coal” or ”national gas” levels, to the ”maximum renewable” levels, and is willing to pay additional costs of 2.2% for that change. It is also implicit that the user is willing to pay more 1.47% to increase from score 2 to 4, and 0.73% to increase from 2 to 3. Finally, the Results sheet contains two plots, as can be seen on Figure 5: the one on the left, showing the overall ranking for the scenarios, and the one on the right showing the contribution of each crite-

9

Figure 4: MCDA tool environment (Excel Sheet 4): Impacts and Criteria Weighting

Figure 5: MCDA tool environment (Excel Sheet 5): Results. Here the user can validate his perceptions. rion. The ranking is scaled so that the best scenario is scored by 100. On the given example, ”coal” scenario is the most rated, while the ”Cost” criterion is assigned as the most important.

4. Results In this section the results are presented. The collaboration with academics took place in two phases. In the first place, the issues that should be included in power planning decision-making were collected with semi-structured interviews constructed over questions raised in the literature. The results of this exploratory research are described in section 2 of this report and published in [3]. In a second phase, the MCDA tool was sent by e-mail to approximately 60 academics, with background in energy, either from Economics or Engineering (Power Systems/Energy/Environment/Mechanical). The eleven experts that proceeded to the evaluation of the scenarios did it in a period of six weeks. Six of them

10

Figure 6: Aggregation of results responded to the tool by themselves, while the other five respondents were aided in a personal interview, which they found helpful and less time-consuming. Table 4 presents the weights assigned by each respondent to each criterion.

Table 4: Criteria weights. Criterion Costs National Industry Energy Independency Employment Visual Impact Noise Local income Diversity of Mix Rate of Dispatchable Power Investment in the Transmission Grid CO2 emissions Land Use Public Health

Respondents E F G 70 100 100 20 25 37 50 100 0 50 50 37 80 50 9 0 50 9 0 75 0 80 100 15 100 50 30

A 50 30 30 30 1 6 0 15 7

B 80 20 70 60 5 2 30 20 40

C 25 50 70 60 50 25 50 60 25

D 80 50 70 50 0 50 10 20 50

15

20

25

10

0

75

5 0 30

60 5 10

60 40 70

50 20 50

0 20 70

90 75 90

H 80 30 30 75 20 10 10 10 20

I 80 25 35 35 15 20 17 12 30

J 80 30 20 20 10 5 5 20 20

K 80 100 100 100 100 30 70 70 50

18

30

35

5

50

27 5 18

30 60 60

40 15 45

0 5 5

100 20 85

Figure 6 aggregates the results, that were normalized for each respondent, so that the highest weight equals 1 and the lowest equals 0. Costs prevailed as the most important criterion, followed by energy dependency, followed by two social concerns: public health and employment. Least important criteria were noise, visual impact, land use and local income.

11

The resulting rankings are presented in Table 5. There are no dominated solutions, which means that no scenario performs always worse than any other scenario. Even in the case that cost is regarded as the most important criterion, the best solution can either be the cheapest or the most expensive: the proof is that ”Coal” and ”Maximum Renewable”, the cheapest and the most expensive scenarios respectively, were the ones that ranked first more times (4 times each). The only scenario that never ranked first, for any respondent, was ”Hydro-Gas”. However, it is a balanced scenario, since it only ranks in the last place twice, while ”Maximum Renewable” and ”Natural Gas” rank in the last position for three respondents’ profiles. On the other hand, ”Base” is the only scenario that never ranked last place, although only ranks first in two respondents. Figure 7 presents the contrast between respondents favorable to ”Coal” and ”Maximum Renewable” scenarios, showing that while the former group clearly places costs high above any other criteria, the latter have five similarly valuated criteria: costs, public health, energy independency, national industry and employment.

Figure 7: Average profile of respondents that chose either ”Coal” or ”Maximum Renewable” as preferred scenarios. The obtained results confirm that costs are still the main obstacle for the incorporation of more renewable energy in electricity systems. Such as [19] case, our scenario ranking was also very sensitive to the input of costs weight. What these results have shown is, in first place, that respondents felt it is important to trade-off costs with other criteria, hence the utility of multi-criteria methodologies. Only on rare occasions did a respondent assign zero to the weight of one criterion, but was free to do it in any criterion he wished to (if he assigned zero to all criteria besides costs, obviously the Coal scenario would be the first in the ranking, since it is the cheapest solution). Secondly, it is the magnitude of the trade-off that induces the divergence in the final rankings. For example, for the second most rated criterion, energy dependency, one respondent suggested that more information should be given when valuating this criterion (”in the worst case for fuel cost projections, how much would the price of the solution increase?”), otherwise it becomes difficult to state how much would value the criterion. However, using more information would significantly increase the response time.

5. Conclusions In this paper, a tool to evaluate scenarios for electricity production was proposed. The tool uses multicriteria decision analysis, and comprises a set of thirteen criteria, ranging from economic concerns,

12

Table 5: Scenario Ranking. Scenario Base Natural Gas Coal Hydro-Gas Maximum Renewable

A 2 5 3 4 1

B 1 5 3 4 2

C 3 4 5 2 1

D 2 5 4 3 1

Respondents E F G 4 2 3 3 4 2 1 1 1 2 5 4 5 3 5

H 1 4 3 5 2

I 4 1 2 3 5

J 2 4 1 3 5

K 2 5 4 3 1

to environmental and social as well as technical issues. The methodology combines an additive value function that aggregates results from direct weighting and trade-off analysis. The proposed tool was used on the particular case of Portugal, based on a set of scenarios for the electric system in 2020. A group of experts from academia, Engineers, Economists related to the energy sector, participated in the evaluation of these scenarios. From the results obtained, most respondents would be willing to increase the costs of power generation if other issues than the economical ones were to be taken into account. This fact alone proves the utility of MCDA. The evaluated scenarios were ranked differently by respondents with different perspectives, what is not unexpected when using multi-criteria methodologies. In fact, only one of the scenarios, ”Hydro-Gas”, was not chosen to be the preferred by any of the eleven respondents. Aggregating the results, cost was considered the most important criterion, even for most respondents whose preferred scenario was ”Maximum Renewable”. Other also important criteria were the rate of dependency on fuel sources, the employment and the public health issues. Depending on the weight assigned to these criteria, the cost loses relative importance and most expensive solutions may rank first. Future work envisages the collection of additional information, increasing the number of experts involved. Also, being the public acceptance of different technologies a fundamental aspect to ensure the success of strategic scenarios, the work is proceeding with the evaluation of public acceptance of different electricity generation technologies.

6. Acknowledgements This work was financed by: the QREN Operational Programme for Competitiveness Factors, the European Union - European Regional Development Fund and National Funds - Portuguese Foundation for Science and Technology, under Project FCOMP-01-0124-FEDER-011377 and Project PestOE/EME/UI0252/2011. Authors wish to thank all the academics that collaborated in the interviews.

References [1] European Comission. 2011 monitoring report of the eu sustainable development strategy, 2011. European Comission. [2] Espen and Løken. Use of multicriteria decision analysis methods for energy planning problems. Renewable and Sustainable Energy Reviews, 11(7):1584 – 1595, 2007. [3] Fernando Ribeiro, Paula Ferreira, and Madalena Ara´ujo. A methodology to address social concerns in electricity planning. In Proceedings of the Dubrovnik Conference on Sustainable Development of Energy, Water and Environment Systems, Dubrovnik, Croatia, 25-29 September 2011., page 321, 2011.

13

[4] INAG. Plano nacional de barragens com elevado potencial hidroel´ectrico. http://pnbeph.inag.pt/np4/np4/?newsId=4&fileName=pnbeph memoria.pdf, 2007. Instituto da ´ Agua, in Portuguese. [5] REN. A energia e´olica em portugal, 2010. Redes Energ´eticas Nacionais, in Portuguese. [6] REN. Informac¸a˜ o mensal, sistema electroprodutor, 2011. Redes Energ´eticas Nacionais, in Portuguese. [7] S´ergio Pereira, Paula Ferreira, and A. Ismael Vaz. Strategic electricity planning decisions. In Proceedings of the Dubrovnik Conference on Sustainable Development of Energy, Water and Environment Systems, Dubrovnik, Croatia, 25-29 September 2011., page 590, 2011. ˜ para as energias renovAveis ´ [8] Rep´ublica Portuguesa. Plano nacional de acC¸Ao ao abrigo da directiva 2009/28/ce, 2009. Rep´ublica Portuguesa. [9] Organization for Economic Co-operation and development/International Energy Agency, editor. Projected Costs of Generating Energy. Organization for Economic Co-operation and development, Paris, France, 2010. [10] Benjamin Hobbs and Peter Meier. Energy decisions and the environment: a guide to the use of multicriteria methods. International series in operations research & management science. Kluwer Academic Publishers, 2000. [11] Jiang-Jiang Wang, You-Yin Jing, Chun-Fa Zhang, and Jun-Hong Zhao. Review on multi-criteria decision analysis aid in sustainable energy decision-making. Renewable and Sustainable Energy Reviews, 13(9):2263 – 2278, 2009. [12] Justin D.K. Bishop, Gehan A.J. Amaratunga, and Cuauhtemoc Rodriguez. Using strong sustainability to optimize electricity generation fuel mixes. Energy Policy, 36(3):971 – 980, 2008. [13] Fernando Ribeiro, Paula Ferreira, and Madalena Ara´ujo. The inclusion of social aspects in power planning. Renewable and Sustainable Energy Reviews, 15(9):4361 – 4369, 2011. [14] Paul Joskow. Apples and oranges: Don´t compare levelized cost of renewables. The Electricity Journal, 23(10):3 – 5, 2010. [15] Andrew Mills, Ryan Wiser, and Kevin Porter. The cost of transmission for wind energy in the united states: A review of transmission planning studies. Renewable and Sustainable Energy Reviews, 16(1):1 – 19, 2012. [16] Athanasios I. Chatzimouratidis and Petros A. Pilavachi. Multicriteria evaluation of power plants impact on the living standard using the analytic hierarchy process. Energy Policy, 36(3):1074 – 1089, 2008. [17] European Commission. External costs: Research results on socio-environmental damages due to electricity and transport, 2003. European Comission. [18] Boris Krey. Scope of electricity efficiency improvement in switzerland until 2035, 2008. [19] G. Heinrich, L. Basson, B. Cohen, M. Howells, and J. Petrie. Ranking and selection of power expansion alternatives for multiple objectives under uncertainty. Energy, 32(12):2350 – 2369, 2007.

14

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

Comparison of sophisticated Life Cycle Impact Assessment Methods for Assessing Environmental Impacts in a LCA Study of Electricity Production Jens Buchgeister Karlsruhe Institute of Technology, Institute for Technology Assessment and Systems Analysis, Karlsruhe, Germany, [email protected]

Abstract: The methodology of life cycle assessment (LCA) has been applied widely to design for environment (DfE ) of energy conversion proc esses. For the identification of weak points within the process in order t o reduce the environmental impacts the selection of the life cycle impact assessment (LCIA ) method has an influenc e on the LCA results. To understand the differences between various LCIA methods a comparative analysis in general structure, extent of considered environmental aspects, and mathematical relationship for quantifying the cause-effect chain from emissions to an impact in the environment is conducted. Furthermore the life cycle assessment is carried out in a case of electricity production by means of solid oxide fuel cell (SOFC) with integrated allothermal biomass gasification using three different LCIA methods: Eco-indicator 99, CML 2001 (baseline), and Impact 2002. It can be present ed that in t his case of electricity production process the supply of biomass and the production of the SOFC have the highest environmental impacts for all applied LCIA methods. But the divergences of the LCIA methods leads to the result that different chemical pollutants cause the highest environmental impact. For the Eco-Indic ator 99 the highest contribution comes from the particles PM 2.5 whereas for CML hydrogen fluoride, for IMPACT 2002 (endpoint) dioxin (TCDD) is responsible for the highest environment al impact. It can be shown depending on the applied LCIA method the highest contribution on the overall environmental impact can be differ. In due to the result the recommendation is made to use more than one LCIA met hod in order to get much more information in detail of environmental pollutants.

Keywords: comparative analysis, electricity production, LCA, life cycle impact assessment

1. Introduction Life cycle assessment (LCA) has been applied widely to design for environment (DfE) of energy conversion processes. For the identification of weak points within the process in order to improve the environmental performance the selection of a life cycle impact assessment (LCIA) method has an influence on the results. This knowledge is included in an ISO standard on LCA [1,2] for the part of mandatory elements in the LCIA. The current ISO standard takes into account existing uncertainties in environmental impact assessment. For this reason have been constructed a few consensus-oriented publications [3,4,5,7] over the years by working groups under the auspices of SETAC describing state of the art and best available practice on LCIA. In further consequence a number of different LCIA methods are implemented in commercial LCA software products, and are available to LCA practitioner. A general survey of LCIA methods is compiled during phase I of the international Life Cycle Initiative, a concerted project between United Nations Environment Programme (UNEP) and SETAC [8]. Furthermore has been pointed out in papers [9,10] which are carried out comparisons of a few different LCIA methods for specific case studies that main differences are existed in toxicological impacts. Therefore in a European project OMNIITOX (Operational Models aNd Information tools for Industrial applications of eco/Toxicological impact 15

assessments) is focused on methods for the characterisation of toxicological effects and data availability in order to improve and harmonise the practicality of toxicological impact assessment in LCA [11]. The results of the OMNIITOX project are integrated in the new designed LCIA methods Impact 2002. The present paper reports a comparison of the three LCIA methods Eco-indicator 99 [12], CML 2001 [13], and Impact 2002 [14] based on a case of electricity production by means of SOFC with integrated allothermal biomass gasification. The goal is to reveal the causes for the differed impact results between the investigated LCIA methods.

2. Methodology The LCA is an internationally established and standardized method for the analysis of the complete life cycle of products and services [1,2]. It analyses the consumption and emission of material flows from all process steps within the life cycle. This means the supply of the input streams, especially fuel, and the full life cycle of components. Inventories of elementary flows (i.e., consumption of natural resources and energy carriers as well as emissions) are compiled following the guidelines of international standard approaches [1,2]. The basis of the inventory result, which are calculated for the investigated life cycle processes, are the general physical laws of conservation of energy and mass. The accuracy of this procedure depends on the assumptions for each modelled process and the entire system. Based on the life cycle inventory (LCI) result the environmental impacts for various environmental impact categories are calculated by a quantitative LCIA method. An impact category describes the impact pathway between the LCI results and their environmental endpoint(s) or so-called areas of protection. It includes a cause-effect chain by using quantitative characterization indicators based on an environmental model (see figure 1). The measuring and modelling of environmental impacts are complex and it is affected by the two different main schools of approaches: endpoint and midpoint LCIA methods [6,7].

Tropospheric ozone depletion

effect mechanism 1

Midpoint

Parti al pressure (mPa) 0 5 10 15 20 25 30

Ozone depleted emission (FCKW)

Effect indicator: Ozone Depletion Potential (ODP)

O3

ODPx

x

O3 CFC

11

Stratospheric ozone depletion

5 Height (km)

Inventory results

Reduction of ozone concentra. Increased UV-B radiation

Trop opa us e

15 20 25

Stratosphere

30 35

Exposure of UV-B radiation

effect mechanism 2 Skin

Eye

Body

Soil plants

Plankton

Skin Immune Terrestrial PhytoCataract cancer system vegetation plankton Endpoint

Troposphere

10

Human health

Biotic natural environment

Plants Crop

Wood

Plastic Material

Biotic & abiotic manmade environment

Fig. 1. Example of impact pathways of stratospheric ozone-depleting substances [adapted from 7]

The terms midpoint and endpoint mean the location of the environmental impact category indicator. In line with international standard the category indicator can be located at any point between the LCI results and the category endpoints [2]. The midpoint approach describes quantitative by the category indicator only the potential modification of the environmental state. In contrast to them the 16

indicator of endpoint approaches characterize based on the measurable modified environmental state to real impacts and modifications on one or more of the four defined areas of protection: human health, biotic and abiotic natural environment, biotic and abiotic natural resources, biotic and abiotic man- made environment. To clarify the mentioned differences between midpoint and endpoint approach figure 1 presents exemplarily the complete impact pathway from the released ozone depleted substances like CFC to the area of protection for the impact category stratospheric ozone layer depletion. The dispersion of ozone-depleting substances takes place in the troposphere. At an average of 4 years the ozone depleting emissions achieve the higher layer of more than 15 km, the stratosphere, in order to accumulate their completely [12]. During the way of ozone depleting substances from the bottom to higher air layer it comes by means of incident solar radiation to the dissociation of ozone depleting substances. Due to the dissociation chlorine and bromine atoms are generated as free radicals which reduce the ozone concentration measurable in the stratosphere. This mechanism the reduction of ozone concentration is the category indicator which is very similar for all ozone depleting substances for example CFC, HCFC or halon. For this reason the quantitative extent of reduced ozone concentration can be calculated between all these ozone depleting substances as relative reference. For the impact category stratospheric ozone layer depletion the midpoint indicator is characterized by the effect to reduce the ozone concentration (ozone depletion potential, ODP) in the stratosphere. Furthermore all released substances, which has an ODP characterized with regard to their quantitative impact in relation to the reference substance CFC-11 (trichlorfluormethane) in kg CFC-11 for stratospheric ozone depletion (on the left side of figure 1). Due to the described effect mechanism of reduced ozone concentration in the stratosphere the solar radiation on the earth is increasing. The resulting effect is an permanent increasing of ultraviolet B (UV-B) radiation on the earth surface [20]. As rule of thumb is defined that per percent reduced ozone concentration in the stratosphere the UV-B radiation is increased nearly about 2 percent. The increasing UV-B radiation consequently leads to an increase of skin cancer for humans. The direct interrelation between UV-B radiation and different types of skin cancer is verified observed in literature [21]. In the same by health studies is known that a longer exposition of higher UV-B radiation leads for the human eye to cataract. Moreover an increasing of UV-B radiation has also an harmful impact on the plants and animals because they have also no protective mechanisms like the human. But at the moment it is not possible to quantify these impacts on plants and animals as well as for materials like plastics when it is used as building material. That means for endpoint approaches can only quantitative operationalised the impact interrelation between human skin cancer and cataract by an continuous increasing of UV-B-radiation. For those impacts is used the concept of disability-adjusted life years (DALY) which calculate the losses of life years or an equivalent in the case of disease effects. A comprehensive description of the DALYconcept is shown in following literature [22,23].

2.1. Conclusion of comparison between midpoint and endpoint approach In comparison between midpoint and endpoint approach to quantify the environmental impact interrelation within an impact category is pointed out that the endpoint approach needs a longer cause effect chain from the emitted emissions to the area of protections (endpoint categories) than midpoint approaches. Thereby the environmental impact is differentiated viewable on each area of protection, but currently it is not possible the specific changes within the lithosphere, hydrosphere, atmosphere and biosphere to determine in its entirety. Because there is lack of knowledge and data in order to describe the impact pathways of emission correct differentiated in space respectively the effort for data collection is too high. Furthermore leads the long cause effect chain within an endpoint approach to greater uncertainty with each additional chain link. Due to the fact that 17

midpoint approach currently has a lower uncertainty of mathematical function to quantify the environmental impact it have to preferable used. As compared LCIA methods is chosen the CML 2001 method which is a classical agent of midpoint approach, the Eco-indicator 99 as agent of endpoint approach and the IMPACT 2002 method which implements both a midpoint and endpoint approach. A structural comparison of different used method is described in chapter 2.5.

2.2. Life cycle impact assessment method – CML 2001 The CML 2001 method is a classical agent of a midpoint life cycle impact assessment method. It is developed by the institute of environmental science (CML) of Leiden University. The CML has a long tradition in the methodological development of LCIA [13]. The general structure and used impact categories of the CML 2001 method (baseline) is presented in figure 2. Inventory Minerals a. fossil energy carrier Land occupation Emissions NOx, SOx, PO4

Abiotic resource depletion

Res. depletion indicat. - ref. kg Antimon

Land competition

Land use indicator - reference 1 m²

Spezifiic contribution in relation to

Acidification

H+ - concentration - reference kg SO2

Eutrophication

Eutrophication potential - ref. kg PO4

total

Pesticides

Ecotoxicity – terrestrial, marine aquatic and freshwater aquatic

Ecotoxic substance stress in water and soil - reference kg 1,4-Dichlorbenzol

emissions

CO2, CH4

Climate change

Global warming potential - ref. kg CO2

of Western

HFCKW

Stratospheric ozone layer depletion

Ozone depletion potential - ref. kg R11

Europe

Particle Matter Heavy metals

Human toxicity

Effects of toxic substances on humans - ref. kg 1,4-Dichlorbenzol

per yearr

NMVOC

Photochemical oxidant formation

Ozone ref. kg Ethen Ozone formation formation potential index – ref.-kg ethen

Classification

Midpoint Characterisation

Normalization each category

Cumulated Environm.I mpact [ year]

Weighting (each c ategory =1 ) (for comp ara tive analysis)

Fig. 2. General structure and used impact categories of the LCIA method CML 2001(baseline) [13]

Each impact category is characterized by an midpoint indicator which uses a defined reference substance in order to quantify the impact of a classified emission in relation to the reference substance. Usually the CML method is finished after the normalization of each impact category whereby the result shows an environmental profile of different 11 baseline impact categories (see table 1). The step of normalization calculate the specific magnitude of impact category result of the investigated system in relation to a reference information. In the case of CML 2001 method as spatial reference value the total emission of Western Europe per year is selected. Subsequently the normalized result of an impact category has the unit year. To compare the life cycle impact assessment method CML 2001 with endpoint methods in order to quantify the environmental impact an aggregation (step of weighting) of each impact category has to integrate in the model structure. For this reason each impact category was equally weighted and can be cumulative added to the total environmental impact.

2.3 Life cycle impact assessment method – Eco-indicator 99 It is especially developed as endpoint life cycle impact assessment method to support decisionmaking in a design for the environment [12]. Therefore based on the three areas of protection (endpoint categories): human health, biotic and abiotic natural environment, biotic and abiotic 18

natural resources an aggregated value to a single score is created. The aggregation depends on a concept of different socio-culture perspectives which is developed by Hofstetter [23]. From the method developers are recommended a default weighting factor of 2/2/1 between the different endpoint categories human health, natural environment and resources [12,23]. The Eco-indicator 99 was the first developed LCIA method which calculates the environmental burden based on measurable real impacts and modifications on the three areas of protection. For this reason new concepts of environmental models like disability-adjusted life years (DALY) which calculate the losses of life years or an equivalent in the case of disease effects is applied [22,23].The structure and the considered environmental aspects are displayed in figure 3. Abiotic re source depletion - Conc. metallic minerals

Surplus ene energy for future Surplus rgy for extraction future extraction

Availability of fossil primary energy ca rrier

Surplus energy for S urplus energy for future extrac future tion ex traction

Land trans form ation and occupation

Tra nsformation - habitat los s a. popula tion reduct.

Regional effe cts on vascular plant typs

Emissions

Acidific ation/Eutrophic at. pH-conc. /nutrie nt c hanges

Acidity Acidific./Eutroph. /Eutrophic ation eff ects on plants sources (occurre nce ta rget s pecie s)

Pe sticides

Ecotoxic ity of urba n and agriculture used soil

Ecotoxic effects of potential affecte d s pecie s

CO2, CH4

Climate Chnage - Conc. gree nhouse gases

Effects on diseas e

HFCKW

Stratos. ozone laye r deplet. - Conc. ozone deplet. gases

Effects on skin ca ncer and cataract

Inventory Minerals a. fossil energy carrier

NOx, SOx PO4

Local effects on vascular plant typs

Ionization radiation - Conc. radionuclide s

Radionuklide

Effe cts on ca ncer cases a. diffe rent typs of cancer

Pa rticle s, NMVOC

Res piratory effects - Conc . particle m atte r and VOC

Effects on cancer cas es a. diffe rent typs of cancer

He av y metal

Carcinogenes is - Conc. in a ir, wa ter, food

Effects on cancer ca se s a . different typs of cancer

Classification

Fate of substances and effect analysis

Abiotic Res ources [MJ surplus energy] Safe guard of a biotic and biotic natural environment [% spec ies of pla nts ty ps * km ² * Ja hr]

Ecoindicator points

Safeguard of human health

[disabilit yadjusted life years DALY]

Endpoint effect analysis

Normalization and Weighting

Fig. 3. General structure and used impact categories of the LCIA method Eco-indicator99 [12]

2.4 Life cycle impact assessment method – IMPACT 2002 The life cycle impact assessment method IMPACT 2002 proposes an implementation of both midpoint and endpoint approach, linking all types of inventory results via 11 midpoint categories to 4 endpoint categories: human health, abiotic and biotic natural environment, abiotic resources, and climate change [14]. The general structure and the integrated impact categories of the LCIA method IMPACT 2002 is shown in figure 4. In difference to other endpoint LCIA methods the impact category climate change is defined as additional area of protection and therefore it is integrated as further endpoint category. The increasing relevance of the climate change is derived from the interrelation to life support functions of plants, animals, and humans. For the comparative analysis of different LCIA methods the endpoint approach of IMPACT 2002 is selected. Therefore a weighting of the four endpoint categories have to carried out which is included in the model structure of figure 4. Similar to CML 2001 each damage category was equally weighted. The result of IMPACT 2002 is expressed in points.

19

Inventory Inventory Minerals a. Minerals and ossil energy ff ossil fuels carrier

Abiotic resource depletion Concentration minerals - Conc. minerals

Surplus energy energyfor for future Surplus extraction future ext raction

Non-renewable energy carrier – upper heating value

Non-renew able Total Total primary nonprimary energy demand renewable energy

Land Land Use Occupation occupat ion

Safeguard Damage abiotic to Resources resources [MJ] [MJ Energy]

Pesticides Pesticides

Ecotoxicity - water a.Aquatic soil Ecotox. - Terrestrial, Ref.kg kgtriethylene Triethylenglykol --ref. glycol

Ecotoxic effects on and Toxic effects in soil plant water sources

NOx, SOx PO4

Acidification/Eutrophication Terrestrial Acidification/ of soil – ref. kg SO2kg SO2 Nutrification – ref.

Ef Acidifi./Eutrophic. fects on occurrence effects of on plant sources target plant species

Safeguard Damage of abiotic and to biot. Ecosystem natural environment Quality [% Pflanzenplant [% arten species * km² * * 2 Jahr] km * year]

CO2, CH4

Climat Change Climate Change reference kg CO2 -- ref. kg CO2

Effects on lif e support system plant, human etc. function

kgequival. kg CO2 eq CO2

RadioNMVOC nuclides

Phot Ionization ochemical radiation Oxidation -- ref. reference kg ethylene Bq C-14

Respiratory effectscases from Effect s on cancer organics a. different ontyps humans of cancer

HCFC, CFC HFCKW

Stratosphere ozone layer Ozone layer depletion depletion - ref. kg R11 ref. kg CFC-11

Effectsofon skin cancer Effect cancer and andaract cataract cat on humans

Nuclides NMVOC

Photochemical oxidant formation - ref. kg Ethen

Effekt auf Atembeschwerden durch org. Schadstoffe

Particle P article mat ter m atter

Respiratory effects – ref. kg PM 2.5

Eff ect on respiratory by means of inorganics

Heavy metal metal Heavy

Human Human toxicity toxicity - ref. kg Chlorethen chloroethylene

Carcinogenic effects of Subst ances toxic stress toxic substances on humans

Emissions Emissio ns:

Local effect Local effects onon vascular plant species vascular plants

Classification

M idpoint effect analysis

Endpoint effect anal ysi s

Saf eguard Human Damage to Health Human Health [Disability [Verlust an adjusted LebensLife Years jahren (DALY)] DALY]

Impact 2002 poin ts C umu lat ion f or c omp arison o f m etho ds

Normalization of each endpoint category and weighting each cat. =1

Fig. 4. General structure and used impact categories of the LCIA method IMPACT 2002 [14]

2.5 Structural comparative analysis of applied LCIA methods In table 1 is shown a comparison of applied impact categories within the LCIA methods CML 2001 baseline, Eco-indicator 99, IMPACT 2002. The result is that the Eco-indicator 99 and IMPACT 2002 have a similar structure. In difference to the Eco-indicator 99 is the category land transformation not included as well as for the CML 2001 method. Table 1. Comparative Analysis of applied impact categories in the LCIA methods CML 2001 baseline, Eco-indicator 99, IMPACT 2002 Impact categories Abiotic resources Land occupation Land transformation Acidification and eutrophication of soil Climate change Stratospheric ozone layer depletion Photochemical oxidant formation Ecotoxicity in soil Ecotoxicity in marine aquatic water Ecotoxicity in freshwater aquatic Ionization radiation Human toxicity Respiratory effects (by inorganic)

CML 2001 X X X (two categ.) X X X X X X X (human tox.)

Eco-indicator 99 X X X X X X X X X X X X

IMPACT 2002 X X X X X X X X X X X

Furthermore IMPACT 2002 has developed new environmental models for human toxicity and ecotoxicity and for non-renewable energy carriers the total primary energy content is accounted. All of these differences to the Eco-indicator 99 method are marked by hatched areas in figure 4. 20

The greatest structural distinctions in regard to the included environmental aspects between CML 2001 and the others methods are for the categories human toxicity and ecotoxicity. Here is the point that each method use their own environmental model to calculate the dispersion of toxic substances in the different environmental compartments [12,13,14]

3. LCA case study of electricity production For the application of an comparative analysis of presented different LCIA methods a LCA of a thermochemical process for the conversion of biomass to electricity was carried out. The system boundary of the LCA covers all system components (see figure 5) and their respective life cycles, as well as all input streams (biomass, electricity, water) to the overall energy conversion system. The details of the process can be found in [16]. The modelling of a similar process has been reported in [17,18]. Figure 5 shows the flowchart of the process designed for electricity generation of 1 MW alternating current. In the same figure, temperatures, pressures and mol flow rates of the material streams are presented. These values were obtained by modelling and simulation of the process described below. Biomass (wood chips) is fed to an allothermal fluidized bed gasifier (Gasifier) that is heated by an integrated burner. The flue gas of the solid oxide fuel cell (SOFC), which contains non-depleted fuel, represents the feedstock for the burner. The gasification agent is steam generated within the process. At 750°C the biomass is converted to a raw gas which mainly consists of H2, CO, CO2 and CH4. After leaving the gasifier, the raw gas enters the hot gas cleaning facility at 650°C. First it passes through a ceramic particle filter and an adsorber. Char, bed material and ash are removed in the first component, halogen and sulphur compounds are removed in the latter. Steam pulses periodically clean the particle filter. After the adsorber, steam is added to the gas in a mixer (MIX) to adjust the steam-to- gas ratio to a value of 2.5, which is necessary for tar and methane reforming and for preventing coke formation. Before the gas enters the tar reformer it has to be heated from 470 to 900°C to enable the reforming reaction to take place. This is realized in a heat exchanger (HX G4) by transferring heat from the hot anode flue gas (1000°C) from the SOFC to the tar laden gas. The tar is completely reformed to lower hydrocarbons in the catalytic tar reformer (Tar Reform.). G 11

751 1.02 163.6

Biomass B0

G2

Gasifier

25 1.00 28.3

550 1.07 40.9

G1

450 1.06 40.9

650 1.10 40.9

G3B

MIX Adsorber

STH0

641 1.03 79.6

STH3

Electricity

Tar Reform.

220 1.03 302.4

800 1.04 317.4

SOFC Exhaust Air

Electricity

A3 1000 1.04 302.4

SOFC

W P1 1.087 MW

A0B

ST 2

ST 1

25 5,00 36,4

G 13

235 1,00 163,6

Gasifier Flue Gas

Inverter

220 1.03 211.7

=

WP2

1.000 MW

A 4C

501 4,90 36,4

HXST

HX A1

31 1.05 317.4

Syst em Power

WA 0 0.01 7 MW

Stream Label A0 T (°C)l 25 P (bar) 1,0 0 SOFC Exhaust Air Flow (kmol/h) 31 7,4 Air Electricity

1 MW el. (AC)

Biomass Air / SOFC Exhaust Air Water / Steam Gasifier Flue Gas Raw a.Prduct gas /Anode Flue Ga s Electricity

Fig. 5. Flowchart of electricity production by means of biomass conversion process

21

~

Legend:

Blower

Steam

0,0003 MW

WG5 0,033 MW

A4

501 4,90 29,5

220 1.03 90.7

Water WST1

G7 1000 1.03 79.6

Heat G6

759 1.04 71.6

HX G4

Anode Flue Gas

A1

615 1.01 163.6

Pump

25 1,00 36,4

G6 800 1.03 71.6

G 4B

G 12

501 4.90 6.9

W1

G5 759 1.04 71.6

641 1,02 79,6

A4B

HX A5

900 1.05 70.4

G8

G9

A5 524 1,02 90.7

G4

470 1.06 70.4

Raw Gas

Particle Filter

Product Gas

G3

The clean gas is heated to 800°C in an electric heater (Heat G6) before entering the anode side of the SOFC. Preheated air (800°C) is supplied to the SOFC at the cathode. At an operation temperature of 1000°C the fuel cell produces direct current by oxidizing hydrogen and carbon monoxide. Prior to that, methane was internally converted with steam to hydrogen and carbon monoxide. The fuel utilization factor in the SOFC is 69 %. The inverter (Inverter) converts the direct current to alternating current. Ambient air is fed to the air preheater (HX A1) by an electric blower (Blower). Heat from the fuel cell exhaust air (1000°C) is transferred to the outside air. The exhaust air from the SOFC is partly released to the environment and partly preheated in another heat exchanger (HX A5) to about 520°C and fed to the burner, which is integrated into the gasifier. The hot stream in the heat exchanger is the flue gas from the burner that has previously heated the gasifier. The water supply of the system is provided by a pump (Pump) that pressurizes water to 5 bar. Following this, steam is generated from the water in a heat exchanger (HX ST) that transfers heat from the flue gas of the burner integrated into the gasifier. The gasifier model is based on a mass and energy balance and on the reforming reactions taking place in the gasifier. The fuel cell model was been adapted from a model for a tubular SOFC published in [17,18]. Depending on gas composition and operating conditions, the power output as well as the conditions and compositions of the exiting material streams of the fuel cell can be simulated. The LCA of the system is modelled in the material flow software Umberto (version 5.5 [19]). For the case study the assumption was made that no heat demand exists outside the process in this theoretical model to simplify the calculations. Therefore, the product from the overall process is only the generated electricity. As functional unit for the LCA the production of 100 MWh of electricity has been chosen. Furthermore is assumed that the bioenergy facility investigated is situated in central Europe. All other important assumptions of parameters for the LCA are presented in table 2. Table 2. Defined parameters and assumptions for the LCA of the case study of electricity production Parameter Value Functional unit

100 MWh electricity

Lifespan of total plant

100,000 operation hours, respectively 15 years

Lifespan of solid oxide fuel cell heating value of biomass

40.000 operation hours 19.8 MJ/kg

Density of biomass (dry)

169 kg/m³

Moisture of biomass during transport Biomass capacity (wood chips)

40 % 500 kg/h

Transport distance of biomass

50 km

The whole LCI of the conversion process of biomass to electricity was presented in detail in [16].

3.1 Results of the compared LCIA methods Figure 6 presents the percentage of environmental impact of all input streams and system components in relation to the total environmental impact associated with the final product for all applied LCIA methods. It shows that the supply of biomass has the highest environmental impact for all methods but the values varies from 58.7 % (Eco-indicator 99) to 45.0 % (IMPACT) and to 33.5 % (CML).

22

Fig. 6. Percentage of total environmental impact of input streams and system components

One reason for the highest Impact of biomass supply of all used LCIA methods can be explained by the high effort for the transport of the biomass that is transported about a distance of 50 km. Additionally the impact of land occupation especially land transformation which is only included in the Eco-indicator 99 method contributes to the highest value. On the other hand the values for the electrical supply of the blower and heater G6 are clearly higher for the CML method during both others methods show similar values. The reason for this result is the high impact of hydrogen fluoride (HF) by the CML method. This pollutant is produced especially in fossil power plants which have the main contribution in the used EU electricity mix. But the presented consistent results of the applied LCIA methods are not sufficient in regard to the same environmental optimization potential. For this purpose is a dominance analysis necessary in order to find out which elementary flows (emission, resource or land use) are the main contributors on the total environmental impact.

3.2 Comparative Dominance Analysis Figure 7 presents the results of the dominance analysis for the different applied LCIA methods. For the Eco- indicator has the highest contribution the particulate matter (PM 2.5) followed by land occupation and transformation and the demand of crude oil. During for the IMPACT method dioxin emissions (equivalent of 2,3,7,8-tetrachlorodibenzodioxin) are with over 70% responsible for the highest impact. The analysis of CML method shows that hydrogen fluoride and heavy metal emissions in water each with over 30 % exhibit the main load. The variation of main contributors in each case of applied method shows that it depend on the impact categories human toxicity and ecotoxicity which different effect models has implemented. At this crucial point recently published studies are started further development to reduce the uncertainty in these impact categories [15].

23

90%

SOX in a ir

80%

CO2 in a ir

70%

As in wa ter

60%

Natura l gas in ground

50%

Cd in w ater

40%

NOx in a ir

30%

Crude oil in ground

20%

Land Occup. & Tra nsf.

10%

Particles in air

100%

All others NOx in air Particles < 2.5 in a ir

80%

CO2 in a ir Oil, coal, gas, uran Diox in (TCDD) in a ir

60%

40%

20%

Eco-indicator 99

90% 80% 70% 60%

All othe rs Diox in (TCDD) in air SOx in a ir NOx in a ir CO2 in air

50% Oil, coa l, gas 40% La nd Occupation 30% Ni,V ,Co,Cu,Hg in wa ter 20% 10%

HF in a ir

0%

0%

0%

100%

Environmental Impact

All others

E nvironmental Impact

Environmental Impact

100%

IMPACT 2002

CML 2001

Fig. 7. Comparative dominance analysis of applied LCIA methods Eco-indicator 99, IMPACT 2002 and CML 2001

4. Conclusion It is shown that the different life cycle impact assessment methods exhibits structural distinctions in regard to the included environmental aspects as well as for the mathematical function to quantify the environmental impact. Today neither a midpoint nor a endpoint method is available which all quantitative measurable environmental aspects could be involved. For this reason it makes sense for an analysis of environmental loads using more than one method in order to collect the environmental aspects in their broadness. Due to the fact that midpoint approach has a lower uncertainty of mathematical function to quantify the environmental impact it have to preferable used. In the case of electricity production by means of biomass gasification with high-temperature solid oxide fuel cell is shown that all used LCIA methods provide similar results in respect of the components with the highest environmental impacts. The relevant components are the biomass supply, gasifier, SOFC, and the electricity supply of heater G6. However, the dominance analysis shows that as a function of the used LCIA method different inventory results (elementary flows) are responsible for the main contribution. All of these elementary flows are assigned to the impact categories human toxicity and ecotoxicity whose toxicity characterization models are different for each used LCIA methods. In the modelling of human toxicity and ecotoxicity have the LCA respectively LCIA method their greatest uncertainty. At this crucial point recently published studies are started further development to reduce the uncertainty in these impact categories [15]. In due to the result the recommendation is made to use more than one LCIA method in order to get much more information in detail of environmental pollutants.

References [1] International Organization for Standardization (ISO), Environmental Management - LCA Principles and Framework, European Standard EN ISO 14040. Geneva, Switzerland, 2006 [2] International Organization for Standardization (ISO), Environmental Management - LCA Requirements and Guidelines, European Standard EN ISO 14044. Geneva, Switzerland, 2006 [3] Udo de Haes H.A. (ed.), Towards a methodology for Life Cycle Impact Assessment. Society o f Environmental Toxicology and Chemistry (SETAC-Europe), Bruxelles, 1996 [4] Udo de Haes H.A., Jolliet O., Finnveden G., Hauschild M.Z., Krewitt W., Müller-Wenk R., Best Available Practice Regarding Impact Categories and Category Indicators in LCIA. 24

Background Document for the Second Working Group on LCA of SETAC-Europe (WIA-2), Part 1: Int. Journal of LCA 1999;4(2):66-74, Part 2: Int. Journal of LCA 1999;4(3):167-174 [5] Udo de Haes H.A., Jolliet O., Finnveden G., Goedkoop M., Hauschild M.Z., Hertwich E., Hofstetter P., Klöpffer W., Krewitt W., Lindeijer E., Müller-Wenk R., Olsen S., Pennington J., Steen B. (eds), Life Cycle Impact Assessment – Striving towards best practice. Society o f Environmental Toxicology and Chemistry (SETAC), Pensacola, USA, 2002 [6] Bare J., Hofstetter P., Pennington D., Udo de Haes H.A., Midpoint versus Endpoint: The Sacrifices and Benefits - Life Cycle Impact Assessment Workshop Summary. Int. Journal o f LCA 2000;5(6):319 - 326 [7] Jolliet O., Müller-Wenk R., Bare J., Brent A., Goedkoop M., Heijungs R., Itsubo N., Peña C., Pennington D., Potting J., The LCIA Midpoint-Damage Framework of the UNEP/SETAC Life Cycle Initiative, Int. Journal of LCA 2004;9(6):394 – 404 [8] Jolliet O., Dubreuil A., Gloria T., Hauschild M.Z., Progresses in Life Cycle Impact Assessment within the UNEP/SETAC Life Cycle Initiative. Int. Journal of LCA 2005;10(6):447-448 [9] Dreyer L.C., Niemann A.L., Hauschild M.Z., Comparison of Three Different LCIA Methods : EDIP97, CML2001 and Eco-indicator 99. Int. Journal of LCA 2003;8(4):191-200 [10] Brent A., Hietkamp S., Comparative Evaluation of Life Cycle Impact Assessment Methods with South African Case Study. Int. Journal of LCA 2003;8(1):27-38 [11] Pennington D. (ed.), OMNIITOX: Operational Models aNd Information tools for Industrial applications of eco/TOXicological impact assessment. Int. Journal of LCA 2004;9(5):281-342 [12] Goedkoop M., Spriensma R., The Eco-indicator 99: A damage oriented method for life cycle impact assessment. Methodology Report, Amersfoort, Netherlands, 2000, Available at: www.pre.nl [13] Guinée J.B. (ed.), LCA: An Operational Guide to the ISO Standards. LCA in Perspective; Guide; Operational Annex to Guide. Centre for Environmental Science, Leiden University, Netherlands, 2001, http://www.leidenuniv.nl/interfac/cml/ssp/databases/cmlia/index.html [14] Jolliet O., Margni M., Charles R., Humbert S., Payet J., Rebitzer G., Rosenbaum R.K., IMPACT 2002+: A New Life Cycle Impact Assessment Methodology. Int. Journal of LCA 2003;8(6):324-330 [15] Goedkoop M., Heijungs R., Huijbregts M., De Schryver A., Struijs J., van Zelm R., ReCiPe 2008 - A life cycle impact assessment method which comprises harmonised category indicators at the midpoint and the endpoint level. Report I: Characterisation, Ministry of Housing, Spatia l Planning and Environment (VROM), Netherlands, 2009, Available at: www.lcia-recipe.net [16] Meyer L., Exergiebasierte Untersuchung der Entstehung von Umweltbelastungen in Energieumwandlungsprozessen auf Komponentenebene: Exergoökologische Analyse [PhD]. Darmstadt, Germany: University of Darmstadt 2006. [17] Panopoulos K. D. et al., High temperature solid oxide fuel cell integrated with novel allothermal biomass gasification: Part I: Modelling and feasibility study, Journal of Power Sources 2006;159(1), 570-585 [18] Panopoulos K. D. et al., High temperature solid oxide fuel cell integrated with novel allothermal biomass gasification: Part II: Exergy Analysis, Journal of Power Sources 2006;159(1), 586-594 [19] ifu Hamburg, ifeu Heidelberg, editor, Umberto –Software for Material and Energy Flow Analysis and Life Cycle Assessment. User handbook Part 1 and 2, Version 5, Hamburg, Germany, 2011 (in German) [20] World Meteorological Organization (WMO), Scientific Assessment of Ozone Depletion: 1998. Executive summary. Global Ozone Research and Monitoring Project - Report No. 44. Geneva, Switzerland, 1999 25

[21] International Agency for research on cancer (IARC), Solar and Ultraviolet Radiation. Monographs on the Evaluation of carcinogenic Risks to Humans, Vol.55, Lyon, France, 1992. [22] Murray Chr., Lopez A., The Global Burden of Disease. World Health Organisation (WHO), World Bank and Harvard School of Public Health, Boston, USA, 1996 [23] Hofstetter P., Perspectives in Life Cycle Impact Assessment. A structured Approach to Combine Models of the Technosphere, Ecosphere and Valuesphere. Kluwers Academic Publisher, Dordrecht, Netherlands, 1998

26

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

Defossilisation assessment of biodiesel life cycle production using the ExROI indicator Emilio Font de Moraa , César Torresb , Antonio Valeroc and David Zambranad a

Centre of Research for Energy Resources and Consumption – CIRCE, Universidad de Zaragoza, Mariano Esquillor, 15, 50018, Zaragoza (Spain), [email protected], CA

b

Centre of Research for Energy Resources and Consumption – CIRCE, Universidad de Zaragoza, Mariano Esquillor, 15, 50018, Zaragoza (Spain), [email protected]

c

Centre of Research for Energy Resources and Consumption – CIRCE, Universidad de Zaragoza, Mariano Esquillor, 15, 50018, Zaragoza (Spain), [email protected] d Centre of Research for Energy Resources and Consumption – CIRCE, Universidad de Zaragoza, Mariano Esquillor, 15, 50018, Zaragoza (Spain), [email protected]

Abstract: Ensuring the s ustainability of biofuel´s life cycle production has become a mandatory requisite for the European Union´s Member States if they want to account the biofuels consumed in their territory for the established consumption binding 2020 targets. The E U Renewable Energy Directive puts the focus of sustainability on the reduction of greenhouse gas (GHG) emissions and the protection of high biodiversity lands and high carbon stock lands. The current legislative framework does not take into account the consumption of non-renewable resources. Starting from the paper “Assessment of biodiesel energy sustainability using the ExROI concept” published in Energy, which defined the ExROI (Exergy Return on Investment) indicator and applied it to the well-to-tank production processes of biodiesel from rapeseed, sunflower and palm oils, this paper proposes: First, extend the use of the ExROI c oncept, which involves exergy cost accounting, to the life cycle analysis i.e. adding the processes of the production of inputs used in the biodiesel chain; Second, extend the calculations to soybean and used c ooking oil based biodies els; Third, assessing ways to “defossilise” the production cycles to decrease their non-renewable sources consumption. This paper demonstrates that the ExROI indicator is a better indic ator than the EROI (Energy Return on Investment) which only considers energy flows. Also, it shows that the life cycle production of the biodiesels considered has positive ExROI values and that the ExROI value can be improved up to 26.51 taking into account several relatively simple improvements actions. This means that for one exergy unit of non-renewable sources invested in the process more than 26 units of biodiesel can be obtained. This value evidences that biodies el can be around five times more sustainable than fossil diesel, from the point of view of the consumption of non-renewable resources, if adequate measures are taken into account.

Keywords: Biodiesel, FAME, Sustainability, Defossilisation, Exergy, Exergy cost, EROI, ExROI.

1. Introduction A portfolio of alternative fuels, covering electricity, hydrogen, biofuels, methane, LPG and others, is necessary to meet the policy objectives of the European Union [1]. The National Renewable Energy Action Plans produced by the EU Member States setting the pathways to achieve the 2020 targets of the Renewable Energy Directive 2009/28/EC (RES Directive) [2], which aims at achieving a 10% use of renewable energies in transport, show that the 85% of the target for renewable energy in transport will come from first generation biofuels, i.e. conventional bioethanol and biodiesel. From these, biodiesel will play a substantial paper being 65.9% of the total target [3]. These biofuels will need to comply with the sustainability criteria set in the RES Directive which relate to the reduction of life cycle GHG emissions compared to fossil fuels, the protection of biodiversity and the exclusion of use of high carbon stock lands. Beyond these criteria, in order to ensure sustainability, it is necessary to defossilise the production life cycle as maximum, which 27

means substituting non-renewable fuel energy sources and derived products used in the process, by renewable energy resources and derived products. Based on the ExROI concept, defined in a previous paper [4], this paper proposes several alternatives to defossilise the biodiesel fuel life cycle. ExROI, Exergy Return on Investment is used to calculate the ratio of non-renewable exergy consumed in the system to the exergy that the biodiesel contains. The less non-renewable exergy consumed, the higher the ExROI value will be. This definition is inversely equivalent to the non-renewable exergy cost which accounts the amount of non-renewable resources required to obtain a product. ExROI

P CPnrs

1

(1)

cPnrs

The ExROI concept as used in this paper conjugates two important factors, life cycle assessment and exergy cost analysis. Life cycle assessment allows taking into account all non-renewable resources required from crop cultivation to the transesterification plant (primary processes), including the production of the required inputs (secondary processes), meanwhile exergy cost analysis permits the correct cost assessment taking into account the energy quality of the production flows. Exergy allows the integration of matter and energy flows in the analysis of production systems using the same concept and units for both. An adequate selection of the boundaries of the system is very important, as in any life cycle analysis, thus depending on the processes included in the analyses the values obtained can vary significantly. Our previous paper [4] only took into account the direct production processes, while in this paper the production processes of the inputs to the direct production stages are also taken into account; this means for example the production of fertilisers used in the cultivation of the energy crops or the production on methanol for the transesterification process. This is a more accurate way of taking into account the exergy costs, as in the previous work, the exergy costs of the inputs entering the direct production processes were assumed to be their exergy values, following the Theory of Exergy Cost [5]. By comparing the results of the previous publication with the results obtained in this one, we will be able to understand the weight that the secondary processes have in the consumption of non-renewable resources. The manufacture of machinery and equipment is not considered as this is neither considered in the sustainability criteria of the RES Directive. This paper analyses the production life cycles of biodiesel from rapeseed, sunflower, palm, soybean and used cooking oil. Data used for carrying the calculations are mainly based on the life cycle assessment study (LCA) carried out by the JRC-EUCAR-CONCAWE consortium [6] (JEC study) which have been used by the European Commission to establish the sustainability criteria of the RES Directive and the Fuel Quality Directive (2009/30/CE) regarding the CO 2 emissions from the cultivation of the raw materials to the production of biofuel. In the cases where information was missing from the published databases, other sources of information have been used. For example, the life cycle of the used cooking oil (UCO) has been obtained from CIEMAT [7]. In the specific case of rapeseed biodiesel, a separate analysis has been produced using the SimaPro programme and EcoInvent database. This exercise will allow comparing the ExROI value obtained for one specific product using two different databases.

2. Short description of the life cycles

The biodiesel production processes vary depending on the resource. The biodiesel fuels based on energy crops, i.e. rapeseed, sunflower, palm and soybean begin by the cultivation process in order to obtain oil seeds, or fresh fruit bunches (FFB) in the case of palm plantations. In this stage, 28

fertilizers, pesticides and energy is consumed in different quantities for each crop. From this, each resource follows different stages that are summarised in the table 1. The processes for rapeseed, sunflower and palm oil were explained more in detail in paper [4]. The case of soybean based biodiesel is similar to the rapeseed and sunflower, but here there is no need for drying, and the transport needs increase since soybean is currently cultivated in South America and transported to Europe, where refining and transesterification take place. The extraction is done by using n- hexane for all crops except for the palm oil. In addition, in the palm oil extraction the energy is obtained by burning palm biomass residues obtaining methane and heat. In this work these flows are considered as valuable co-products. The case of UCO is completely different. UCO is considered a residue which in case of not being used, would need to be disposed in a landfill. Given this, the previous stages before the oil becomes a residue (including the use, for example, in a frying pan) are not considered in the analysis. The life cycle starts by the collection and transport of the residue, and is followed by the recycling where the oil is filtered and decanted in order to separate solid particles and water. Once the oil is refined, it is sent to the transesterification plant. In the transesterification process the after-treatment of biodiesel (FAME washing) and the glycerol refining are treated separately to analyse the influence of glycerol in the ExROI calculation. This is the case for all the resources except for UCO where the glycerol is not refined. Table 1. Direct processes considered in the life cycle analysis No. Rapeseed and Palm process sunflowe r

Soybean

UCO

1

Cultivation

Cultivation

Cultivation

Collection and transport

2

Drying

Road transport and Storage

Road and maritime Transport

Recycling

3

Transport

Extraction

Extraction

Transesterification

4

Extraction

Road transport, Depot and M aritime transport

M aritime transport

5

Refining

Refining

Refining

6

Pretreatment and transesterification

Pretreatment and transesterification

Pretreatment and transesterification

7

FAM E washing

FAM E washing

FAM E washing

8

Glycerol refining

Glycerol refining

Glycerol refining

3. ExROI values

The thermoeconomic model of the biodiesel production processes is represented by the productive diagrams for each of the direct processes considered in the life cycle analysis depicting the exergy flows entering and exiting each of the processes. As example, Figure 1 shows the case of rapeseed biodiesel production. Although this diagram only shows the direct processes, the exergy consumed for the production of the inputs is being accounted. From these diagrams it is possible to obtain the Fuel-Product table. Table 2 represents the Fuel-Product table for rapeseed, where Fi refers to the process i of the productive diagram of Fig. 1. It is worth noticing that, as explained in [4], it has been assumed that the consumption of replenishable natural resources such as rain water, CO2 and solar energy in cultivation do not add exergy to the exergy costs. By doing this, the only irreversibilities taken into account for calculating the exergy costs are the ones provided by the non-renewable materials. 29

From the Fuel-Product tables it is possible to obtain the production costs of each process applying the Theory of Exergy Cost [5]. Concisely, we apply its fundamentals as follows: in a specific stage of the process the exergy cost is distributed to all the products (main product and by-products) of the stage proportionally by their exergy value; there is no exergy cost allocated to the waste produced; and the exergy cost of resources entering into the system is equal to their exergy.

Fig. 1. Productive Diagram of rapeseed biodiesel life cycle production business as usual Table 3 shows the production costs of each process for the rapeseed biodiesel life cycle. It is worth noticing that the non-renewable energy cost of the refined glycerol is higher than the one of the final biodiesel. This means that the ExROI of glycerol would be 1.12. The reason behind this high value is that the exergy of glycerol is low while the non-renewable exergy consumption is considerably high in the purification phase where a lot of energy is consumed in the distillation unit. Table 2. Fuel-Product table for rapeseed biodiesel life cycle production (a) business as usual, (b) applying improvement measures (MJ/kg FAME) (a) F1 F2 F3 F4 F5 F6 F7 F8 F0 Total rs 77.3 77.3 E0 E nrs 0

P1 P2 P3 P4 P5 P6 P7 P8 Sum

12.0

0.6 77.3

0.1

2.9

0.4

4.4

0.2

1.0

77.3 76.5 43.0

11.8 41.3 40.6

89.3

77.9

77.4

79.4

43.5

45.7

40.8

3.1

4.1

40.0 2.5 54.4

20.7 77.3 77.3 76.5 54.8 41.3 43.7 40.0 2.5

As explained above, the inverse of the unit exergy cost of the non-renewable resources of biodiesel gives the ExROI value according to equation (1). Table 4 shows the ExROI values, unit exergy costs (cp ) and non-renewable unit exergy costs of the different biodiesel fuels (cp nrs). The nonrenewable exergy cost and ExROI values are also calculated when using the exergy content of the inputs to the system instead of their exergy costs. The term “Difference” (ExROI with energy ExROI with exergy costs) establishes the influence of the production processes of the inputs on the ExROI value. The term renewability establishes the weight of the renewable exergy costs with respect the total exergy costs. As it can be observed in Table 4, all the biodiesel sources have ExROI values higher than one, which means that for one unit of non-renewable resources used in their production, more than one unit of biodiesel is obtained. The most sustainable one is the biodiesel produced from used cooking 30

oil, followed by palm, sunflower and rapeseed. The less sustainable is the soybean oil one which almost has a 1:1 relation.

Table 3. Non-renewable production costs of rapeseed based biodiesel production in the business as usual scheme and applying improvement measures (CP in MJ/kg FAME) Business as usual Process

Product

cPnrs

CPnrs

Cultivation Drying Transport Extraction Refining Transesterification

Seeds Dried seeds Dried seeds Crude vegetable oil Refined oil Crude biodiesel + crude glycerol Final biodiesel Refined glycerol

0.156 0.163 0.166 0.282 0.305

12.05 12.57 12.71 15.49 12.59

0.387

16.93

0.398 0.890

15.96 2.22

Biodiesel drying Gly. purification

The consideration of the production inputs (secondary processes) that enter the direct production processes plays an important role in the ExROI value as, except for the soybean oil based biodiesel, the ExROI is reduced in almost 2 units when the exergy costs, instead of the exergy values, are taken into account (see term Difference). Table 4 also shows the EROI values of the biodiesel fuels from energy crops. As it can be observed, this indicator shows very similar values for rapeseed, sunflower and palm oil biodiesels, while these fuels have quite different ExROI values. The reason behind this is that EROI value only takes into account the consumption of energy sources while the ExROI values also considers the consumption of mass flows. The rapeseed based biodiesel consumes more mass inputs than palm and sunflower giving as a result a lower ExROI value. These results demonstrate that the ExROI concept is a better indicator of resource sustainability than the EROI concept. The value for soybean is lower and similar to the ExROI value. The ExROI values obtained above could be considered positive news and an objective indication of which biodiesel resources should be primarily promoted. Hall et al. [8] recommend that the minimum EROI society must attain from its energy exploitation to support continued economic activity and social function is about 3:1 and therefore, biodiesel which life cycle provides an ExROI value lower than 3 should introduce measures to improve their values or be discouraged. In addition, the ExROI values of biodiesel fuels should be compared to its direct competitor, i.e. fossil diesel fuel. According to Cleveland [9], the EROI value of gasoline (and therefore of diesel as they are products of the same process) is in the range of 6 to 10. Given this, the life cycle production processes should be defossilised in order to obtain at least the same ExROI values as fossil diesel fuel. Table 4. EROI, ExROI and unit exergy cost (MJ/MJ) for the different types of biodiesel fuels Source Using Exergy Costs Using exergy values Difference EROI Renewability nrs cP cP ExROI cPnrs ExROI Rapeseed 1.81 0.40 2.51 0.23 4.37 1.86 2.68 78 Sunflower 1.67 0.32 3.17 0.21 4.78 1.61 3.10 81 31

Palm Soybean UCO

1.89 1.98 1.42

0.28 0.59 0.22

3.57 1.69 4.54

0.19 0.44 0.16

5.31 2.26 6.44

1.75 0.57 1.91

3.13 1.62 -

85 70 84

4. Sensitivity Analysis As it has been explained above, the results obtained are based on the consumption data of the life cycle assessments produced by the JRC, EUCAR and CONCAWE [6]. These numbers are fixed and based on specific assumptions. However, consumption in biodiesel production can vary significantly depending on many circumstances, for example, temperature and soil conditions at cultivation, carrying distance, quality of the oil at the transesterification plant, etc. This variation influences the exergy costs of the external resources entering the system and therefore the ExROI value of the product. In order to understand the effect that variations in the consumption of external resources may have in the ExROI result, a sensibility analysis is performed. The exergy cost of the product can be calculated by the following equation: t

CP

P* Ce

(2)

Where P* is the product matrix and Ce is the exergy cost of the external resources entering into the system. As the external costs do not depend on the product matrix a variation in the exergy costs can be calculated by: CP If

* ij

t

P*

Ce

(3)

is a generic element in the production matrix we obtain:

CP , i

* ji

Ce , j

(4)

If we express equation (4) in terms of elasticity coefficient, and we only consider the non-renewable exergy costs, we obtain: * % CPnrs, i Cenrs , j ji (5) % Cenrs CPnrs, i ,j Applying this equation to rapeseed biodiesel it can be obtained that a 10% variation of the nonrenewable exergy costs entering the first process (cultivation) results in a variation of 5.5% in the production costs of the biodiesel product. On the other hand, if we calculate the impact with respect the consumption of non-renewable resources following the elasticity coefficient of eq. (6), we would obtain that an improvement on the efficiency of the system of 10% would result in an improvement of the ExROI value of 7.42%. % FTnrs % ki

nrs

CF ,i

(6)

nrs

FT

This value shows that in face of possible variations, errors or deviations in the introduction of nonrenewable resources consumption data into the system, the exergy cost does not vary significantly and, as a result, it indicates that the ExROI value is a consistent indicator.

5. Defossilisation of rapeseed life cycle 32

This section analyses the impact of different actions that could be introduced in the life cycle of rapeseed based biodiesel in order to defossilise the process and obtain higher ExROI values. Given the allocation system applied as defined by the Theory of Exergy Cost, such actions should focus on reducing the consumption of non-renewable resources; decreasing the production of residues; finding a value for the residues, in order to allocate exergy costs to their flows; and reducing the amount of inputs, which reduces the exergy entering the system and therefore the exergy costs of the products. The results of the actions explained in this paragraph are gathered in table 6. Starting from the cultivation unit, the first possible action is the use of organic fertilizers instead of inorganic fertilizers which consume high quantities of non-renewable resources. The organic fertilizer considered is compost which provides NPK, for which mass and energy consumption data have been obtained from the EcoInvent database [10]. In such a case, the ExROI value is improved (ceteris paribus) in 58%, i.e. to 3.97. Another possibility in this stage is the selection of crop varieties with high oil contents. The JEC study considers that rape seeds have a content of oil of 0.405 kg of oil/kg of seed. If we choose a seed with an oil content of 0.445 kg of oil/kg of seed [11], which is a 9.9% higher, an ExROI of 2.56 is obtained, which is a 2% higher that the business as usual case. Finally, if 50% of the straw produced at the site is collected and considered as a co-product of cultivation instead of a residue, the ExROI is increased up to 3.42 which is a 36% higher than the business as usual scenario. This scenario assumes that by taking half of the straw which otherwise would stay on the ground, no additional use of fertilisers to cover the potential soil quality losses are needed. This is an interesting solution to reduce the allocation of non-renewable exergy costs consumed in this stage to the main product without affecting the fertility of the soil [12]. In the oil extraction process, the most interesting action is the production of biogas from the rapeseed meal obtained. If this biogas is used in the cycle in order to reduce the use of fossil fuels the ExROI value is increased in 18%, i.e. up to 2.95. In the transesterification unit there are many actions that can be considered. On the one hand, these relate to the substitution of fossil fuel derived methanol by other resources from renewable origin. This would be the case when methanol produced from wood or bioethanol from wheat, are used. In the first case the ExROI value would be increased by 23% (3.08) and in the second case by 14% (2.86). On the other hand, they relate to the partial reintroduction of FAME into the cycle in order to substitute the use of fossil fuels. This substitution could be direct or indirect. Direct substitution consists in the use of biodiesel instead of diesel or heavy oil in the cultivation, drying and transport processes; an indirect use consists in the use of biodiesel instead of fossil energy in the production of the inputs that enter the direct process of the biodiesel production cycle. In both cases, the ExROI is improved by 10% (2.76) and 4% (2.61), respectively, for the considered quantity of biodiesel retrofitted. On the contrary to what could be considered, the anaerobic digestion of the glycerol of the production cycle in order to produce biogas which would be consequently used in the cycle to substitute fossil sources does not increase the ExROI value, but reduces it in -3% (2.44). This could be caused by the low biogas yield of glycerol. If instead of producing biodiesel, we consider the refined vegetable oil as a biofuel to be used directly in adapted engines, the ExROI value of this product for which no transesterification would be needed, would be 3.28 which is a 31% higher than for biodiesel. This provides an indication of the weight that transesterification has in the biodiesel life cycle production. All these actions applied separately do not provide an ExROI value higher than the EROI of fossil diesel fuel, according to Cleveland [9]. However, if we consider the following actions together, we can obtain an ExROI of 26.51 which is 956% higher than the reference situation and quite higher than fossil diesel fuel: 33

- Use of organic fertilizer instead of inorganic fertilizers; - Conversion of rapeseed meal to biogas an use in the process; - Use of seed varieties with high oil content; - Use of methanol from wood in the transesterification process; - Partial consumption of biodiesel in the cycle. Figure 2 shows the productive diagram of this best case scenario with all the recirculations of exergy considered. Table 5 shows the non-renewable exergy costs for this scheme where the improvements have been implemented.

Fig.2. Productive Diagram of rapeseed biodiesel life cycle production combining several defossilisation actions It is worth noticing that there could be situations that could worsen the ExROI of biodiesel production. For example, in the case that the glycerol produced in the transesterification process could not be sold as a product but instead treated as a residue, due to the saturation of the glycerol market. In this case, the ExROI would be reduced up to 2.33 i.e. a 7% lower than the case of glycerol being a valuable product. To conclude the analysis, it is important to note that there are certain values and conversion factors that have not been homogenised and could vary substantially, having this variation a significant impact in the ExROI value. This could be the case for example of the exergy content of the rapeseed meal considered. In this paper, the exergy content of the meal has been considered to be 7.4 MJ/kg, which is the metabolisable energy according to [13]. This is the useful energy animals can profit of when eating the meal. If instead this value we take 21.1 MJ/kg which is the value assumed by [14] for the allocation of CO 2 emissions and energy consumption in the life cycle analysis, we obtain that for the same conditions, the ExROI value is increased in 25%, i.e. up to 3.14.

34

Table 5. Non-renewable production costs of rapeseed based biodiesel applying improvement measures (CP in MJ/kg FAME) Improved process Process Product cp nrs Cpnrs Cultivation Seeds 0.013 0.92 Drying Dried seeds 0.020 1.38 Transport Dried seeds 0.020 1.39 Extraction Crude vegetable oil 0.030 1.53 Refining Refined oil 0.032 1.33 Transesterification Crude biodiesel + 0.036 1.59 crude glycerol Biodiesel drying Final biodiesel 0.038 1.51 Gly. purification Refined glycerol 0.090 0.22 Finally, if EcoInvent database is used instead of the data used in the JEC study, the ExROI value obtained is equal to 2.99, which is 19% higher. EcoInvent is currently the world leading life cycle inventory data source. Taking into account this database the processes and consumption data considered in the cycle are slightly different than the ones considered in the JEC study [6]. These two last results show the need of standardise the input and conversion factors in the methodology of ExROI calculation. Table 6. ExROI values applying improvement measures and percentage of variation compared to the business as usual scenario Defossilisation options ExROI % increase Business as usual 2.51 0 Use of organic fertilizers of biologic origin 3.97 58 Using plants with high oil content 2.56 2 Use 50% of straw as a useful co-product 3.42 36 Anaerobic digestion of meal to biogas 2.95 18 Using methanol from wood instead of fossil methanol 3.08 23 Using bioethanol from wheat instead of fossil methanol 2.86 14 Partial consumption of FAME in the cycle (direct use) 2.76 10 Partial use of FAME in the production of resources (indirect use) 2.61 4 Anaerobic digestion of glycerol to biogas 2.44 -3 PVO - Pure Vegetable oil 3.28 31 Combination of several improvement actions 26.51 956 Considering glycerol as residue 2.33 -7 Sensibility analysis: meal higher energy value. Source: I.D.A.E 3.14 25 SIMAPRO EcoInvent database 2.99 19

6. Conclusions This paper is a continuation of paper [4] published in Energy. There the ExROI value was defined and applied to biodiesel production processes: rapeseed, sunflower and palm oil based biodiesels. Here, the number of biodiesel production processes is widened, including soybean and used cooking oil, and the boundaries of the system have been amplified to take into account not only the direct or primary processes: cultivation, transport, extraction, refining, transesterification; but also the secondary processes, i.e. the processes to produce the materials and energy used in the primary processes. With this, this paper relates the ExROI and exergy costs calculations with the life cycle analysis discipline. 35

The results show that biodiesel life cycles are sustainable from the point of view of the use of nonrenewable resources, although improvements are necessary as the ExROI value is considered to be low in the business as usual scenario. While these values range from 1.69 for soybean biodiesel to 4.54 for biodiesel from UCO, the EROI value of diesel fossil is estimated by Cleveland to be around 6. Given this, it is considered that ExROI values higher than 6 must be achieved. This paper demonstrates that by conjugating several defossilisation improvements, which deal with the substitution of fossil materials by renewable materials, the recirculation of biodiesel and reduction of inputs, ExROI values of around 27 can be obtained, which means more than quadrupling the EROI value of fossil diesel. These measures should also be studied from an economic, social and environmental point of view in order to certify their viability. This work is currently being performed and will be shown in future publications.

Nomenclature c unit exergy cots C exergy cost (MJ/kg FAME) EROI Energy return on energy investment ExROI Exergy return on exergy investment F Exergy of the fuel FAME Fatty acid methyl ester (biodiesel) FFB Fresh fruit bunches P Exergy of the product Subscripts and superscripts e External resources eq Equivalent nrs Non renewable sources P Product rs Renewable sources * Exergy cost

References [1] CARS 21 High Level Group, Competitiveness and Sustainable Growth of the Automotive Industry in the European Union. Interim Report 2011. Brussels, 2011 [2] Beurskens L.W.M., Hekkenberg M., Vethman P., Renewable Energy Projections as Published in the National Renewable Energy Action Plans of the European Member States, Covering all 27 EU Member States with updates for 20 Member States; 2011. No ECN-E--10-069. [3] Directive 2009/28/EC of the European Parliament and of the Council of 23 April 2009 on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC. Official Journal of the European Unio n 5.6.2009. [4] Font de Mora E., Torres C., Valero A., Assessment of Biodiesel Energy Sustainability Using the Exergy Return on Investment Concept. ECOS 2011: Proceedings of the 24th Internationa l Conference on Efficiency, Cost, Optimization, Simulation, and Environmental Impact o f Energy Systems; 2011 July 4-7; Novi Sad, Serbia. Linija Projekt 481-431. [5] Lozano, M.A., Valero, A., 1993, Theory of the exergetic cost, Energy, 18 (9), pp. 939–960. 36

[6] JEC - Joint Research Centre-EUCAR-CONCAWE collaboration, Well-to-Wheels Analysis o f Future Automotive Fuels and Powertrains in the European Context” Version 3c, 2011. Available at : [accessed 18.1.2012] [7] Lechón Y., Cabal H., De la Rúa C., Lago C., Izquierdo L., Sáez R.M., Fdez. San Miguel M., Análisis de ciclo de vida de combustibles alternativos para el transporte. Fase I. Análisis de ciclo de vida comparativo del biodiesel de cereales y de la gasolina. Ministerio de Educación y Ciencia y Ministerio de Medio Ambiente, 2005. [8] Hall C.A.S., Balogh S., Murphy D.J.R., What is the minimum EROI that a sustainable society must have? Energies 2009, 2, 25-47, doi:10.3390/en20100025. [9] Cleveland CJ. Net energy from the extraction of oil and gas in the United States. Energy 2005;30:769-782. [10] EcoInvent Centre, Dübendorf, Switzerland. http://www.ecoinvent.org/ [11] Life Cycle Assessment applied to first generation biofuels consumed in France. Department of Bioresources. Directorate of Renewable Energies, Energy Networks and Markets, ADEME, 2009. [12] Christa K., Elisabeth W., Biomass report. MixBioPells project, Intelligent Energy Europe programme. [13] Iowa Soybean Association. Comparative Composition of Various Oilseed Meals Available at: < http://www.soymeal.org/table1.htmlhttp://kinetics.nist.gov/ janaf/> [accessed 18.1.2012]. [14] Lechón Y., Herrera I., Lago C., Sánchez López J., Romero Cuadrado L., Evaluación de l balance de gases de efecto invernadero en la producción de biocarburantes. Estudio Técnico PER 2011-2020; Madrid, 2011.

37

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

Design strategy of geothermal plants for water dominant medium-low temperature reservoirs based on sustainability issues Alessandro Franco, Maurizio Vaccaro Department of Energy and Systems Engineering University of Pisa Pisa, Italy e-mail: [email protected]

Abstract A design strategy based on a multidisciplinary approach for a sustainable design of ORC power plants is proposed. The design of a geothermal plant is discussed, with reference to a case study about a geothermal area in Tuscany, in which a geothermal reservoir at 120 °C is estimated to be available at about 500 m of depth. A qualitative model of the reservoir under specific production/reinjection conditions is discussed. Numerical simulation of geothermal reservoirs is considered an important interacting issue, also to synthesize the data and different scenarios studied. Three factors are fundamental: the maximum energy production, in the perspective of a sustainable exploitation strategy (definition of wells depth and siting, fluid rates extracted/reinjected); the chemical characterization of the fluid (to define the minimum reinjection temperature in order to prevent scaling phenomena); the definition of a reinjection strategy (flow rates, number and depths of reinjection wells and distances between them). Key results are the temperature and pressure profiles and stored energy reduction in the reservoir during plant lifetime. A power plant output range of 250 – 500 kW is considered in order to keep the temperature decrease in the reservoir in a range of 5 - 10 °C during the expected life of the plant. The case study can be seen also as a general value example to discuss how the sustainability of the medium-low temperature geothermal resource makes the design strategy of these plants different with respect to the other renewable source based power plants.

Keywords: Geothermal energy, Sustainability, Binary cycle power plants, Medium temperature geothermal field.

1. Introduction The great part of geothermal resources available around the world are water dominated fields, at temperatures under 150 °C and pressures below 15 bar [1]. The total worldwide expected geothermal potential for power production has been estimated being about 200 GW, [2]. The binary cycle technology using Organic Rankine Cycle (ORC) is the most efficient solution for power production from these fields. Some manufacturers (Pratt & Whitney/UTC, Siemens) have proposed small size (0.2 MW) standard machinery and power conversion systems. This can be a key element for a large diffusion of geothermal binary cycle plants. The size and peculiarity of such plants is often different from the industrial practice of power production from renewable energy sources. Sustainability and operational parameters of the resources become fundamental issues for these utilizations, to guarantee a successful productivity of the power unit and maximize the durability. Medium-low enthalpy geothermal resources can be available, in many areas, at relatively low depth (less than 1000 m). This circumstance allows a meaningful reduction of drilling and perforation costs on the total cost of the plant. A simulation of the geothermal reservoir and the time variations of temperature and pressure under exploitation conditions should be carried out before the design of the plant, mainly in case of ORC units utilization. This is heavily important for reservoirs at temperatures below 130 °C. Small size power plants (100 kW – 5 MW) are innovative since in traditional geothermal power industry almost only greater sizes (5-200 MW) have been used (flash systems, dry steam plants). Medium-low temperature resources, to be exploited for power purposes with ORC systems, introduce the possibility of small size units. Project sustainability, durability of 38

the plant and low environmental impact of this technological solution are advantageous key points, although the exploration and characterization phases will assume a huge importance. As their performances are strongly affected by changes in the external parameters (resource, environment), a reliable characterization process needs to be carried out, in order to avoid unacceptable off-design working points. A multidisciplinary approach to geothermal projects is necessary. The interconnections between Geosciences and Energy Engineering backgrounds have been diffusely remarked, in the perspective of geothermal plants development and diffusion [3]. The main task of the geothermal utilization projects is the sustainable utilization of the reservoirs and the maximization of the durability of the resource (mainly in terms of temperature and pressure). For this reason it is very important to consider and analyze the whole “geothermal system” constituted by the power plant, the wells system, the geothermal reservoir and all the links between them and the environment. The key factors governing the optimization process of the design of a plant are the definition of a sustainable mass flow rate extraction (potential assessment) and the reinjection strategy (taking into account the scaling phenomena) [4], [5]. Typical problems due to an incorrect characterization of the resource available are: oversizing of the plant, causing excessive extraction of fluid (the reservoir doesn’t replenish the energy stored); unacceptable scaling phenomena (causing corrosion, productivity drop, net diameter reduction, damaging); excessive cooling of the reservoir or losses of fluid due to wrong reinjection strategy. Numerical simulation of geothermal reservoirs is considered to be a very useful and strategic instrument both for two main tasks in the power production industry: the history matching of the field data (from the exploration/utilization history) and the forecast about exploitation scenarios. Its reliability depends on the accuracy of the input data (a tough problem is the definition of a reliability scale for decision making about production). In this study the design of a geothermal plant is discussed, with reference to a case study: a geothermal area in Tuscany, in which a geothermal reservoir at about 120 °C is estimated to be available (at a depth of about 500 m). The optimal production/reinjection and global design strategy for a small size ORC based power plant is discussed.

2. Low-temperature geothermal technology and binary cycle power plants Medium to low temperature geothermal resources are largely diffused, so that they have become very attractive for electrical power production in the last years. According to the more significant approaches in the literature, the classification of the geothermal resources depends mainly on the temperature value. In Table 1 a classification of the geothermal resources is provided. It is estimated that the major part of the geothermal energy stored worldwide is available at temperatures lower than 150 °C [1-2]. Those resource become particularly interesting when they are available at depths below 1000 m. Table 1. Classification of the geothermal resources depending on temperature [°C]

Low enthalpy Medium enthalpy High enthalpy

Muffler & Cataldi [6]

Hochstein [7]

Benderitter & Cormy [8]

Nicholson [9]

Axelsson & Gunnlaugsson [10]

< 90 °C 90 – 150 °C > 150 °C

< 125 °C 125 – 225 °C > 225 °C

< 100 °C 100 – 200 °C > 200 °C

150 °C > 150 °C

190 °C > 190 °C

39

generator

turbine

G condenser

heat exchanger

cooling device

circulation pump

geothermal reservoir

Fig. 1. Scheme of a ORC unit working on a geothermal reservoir, a doublet of withdrawalreinjection wells is shown Binary power plants based on Organic Rankine Cycle (ORC) are considered to be the best technology to exploit medium-low geothermal resources (mainly water dominated) at temperature below 180 °C. The operating principle of this plant is represented in the scheme of Fig. 1. In a binary cycle power plant the heat of the geothermal fluid rate is transferred to a secondary working fluid (usually an organic fluid with a low boiling point than water at a given temperature) through an heat exchanger. The cooled geothermal fluid is then returned to the ground to recharge the reservoir, while the working fluid expands through a turbine, producing electrical power and then is condensed by exchanging heat with the environment (dry or wet cooling). A geothermal binary power plant is characterized by high brine specific consumption (kg of fluid extracted per unit of power produced) and low First Law efficiencies I (5-10%), even if Second Law efficiencies II are typically in the range 25-45%. They usually require large heat transfer surfaces both for the recovery heat exchanger and for the condensation system. The parasitic power consumption of this auxiliary system is relatively high because of the need for forced ventilation. A dry cooling system can absorb from 10-12% of gross power (under ideal conditions) to as much as 40-50% if the ambient temperature is very close to the condensation temperature. A diffused literature about this technology is available, often based on specific and local industrial application. The efficiencies of binary cycle plants are very sensible to the external thermodynamic parameters ( T of fluid, environmental temperature, fluid pressure, permeability changes), so the characterization of the resource available is a fundamental step. Recently, binary power plants have been installed in Austria and Germany in applications to medium-low temperature geothermal sources (Bertani [11]). Various studies ([12-14]) demonstrates the relevance of the optimization process applied to ORC units, particularly in terms of efficiencies and fluid extraction. These plants often operate through advanced thermodynamic cycles (dual pressure level Rankine cycle or Kalina cycle) and may also use different or unconventional working fluids, such as ammonia-water mixtures (e.g. Husavik, Iceland). The characteristics of some of those plants are given in Franco [15], in which a table collecting a series of available data is present. The temperature range covered is wide (74-124 °C) so that specific brine consumption lies in the range from 44 to 200 kg/s for each MW of electricity produced. The remarkable difference among the various plant performances can be explained in al lot of case because of the different temperature interval ( T=Tgeo-Trej) available between the reservoir and the reinjection temperature. A scheme of temperature values of different plants is given in Fig. 2. Till some years ago each installation was designed for specific conditions at a given location. Every system is generally tailored to specific geothermal fluid characteristics, while the medium and low temperatures applications permit to pursue the perspective of providing “standard machinery”. The possibility of “standard machinery” development is submitted to a proper characterization and 40

potential assessment of each local reservoir, due to the strong dependence on , external parameters (environment/reservoir).

I

and

II

from

T (°C)

160 140

Tgeo Tgeo

120

Trej Trej

100

To To DTeff Teff

80 60 40

M K

Neustadt Glewe (Germany)

we -G G le st ad t N eu

A

lth e im -T

UR

BO D

EN

Altheim (Austria)

AT M Bl um au -O R

M yO R ra d B

Bad Blumau (Austria)

AT

B ad

P i co

V

er m el

ho -O RM

A

T Pico Vermelho (Portugal)

0

Brady Hot Springs (USA)

20

Fig. 2. Temperature values (production – reinjection – environment) in some binary plants.

2.1 Geothermal potential assessment and binary plants A general methodology for the Geothermal Potential Assessment for every kind of geothermal field doesn’t exist, although in literature a lot of different methods and reliable principles have been treated and tested, ([4], [6], [10]). Several studies are about specific geographical areas, more than general value survey and investigation techniques. A lot of instruments are available today to improve the detail of the information needed for the sustainable design of plants. The geothermal potential of a particular area means particularly the definition of temperature (Tgeo) and pressure (pgeo) of the geothermal fluid and of the maximum mass flow rate ( M geo ) that can be extracted maintaining for a long time the thermal properties of the reservoir. A brief list of the general results that should be evaluated is here proposed: thermal energy stored (at a certain time) in the reservoir; temperature, pressure and rate of the extracted fluid; chemical composition of the fluid and saltiness (to define the reinjection temperature Trej); number of wells to be drilled and mutual distances between them; time interval to have an appreciable decrease in the rate and temperature of fluid (or productivity); definition of both the “Base” resource available and of the “Effective” resource, which is useful under favourable and sustainable (economicenvironmental-technological) conditions, [6]; siting of the reinjection wells (number, mutual distances and interference effects); reinjection strategy (effects of reinjection on productivity); number of wells for compensation. The total energy available is the portion extractable and favorably useful of the stored energy in the reservoir. It can be defined roughly multiplying the total energy stored for an appropriate recovery factor (R), [6]. The estimation of the recovery factor is not a trivial task, it depends from the site characterization and it can also be a result of assessment according to the individuals experience. Let us now consider a water dominated geothermal field at moderate temperature. The energy potential of a geothermal reservoir can be first of all referred to the available temperature of the aquifer (Tgeo). In the case of water dominated geothermal fields, the energy available can be referred to an equivalent specific thermal capacity of the reservoir (both rocks and fluid), namely (Tgeo T0 ) , which can be defined referring to the cooling down to a low temperature level, in this case the environmental temperature T0. 41

Fig. 3. Sketch of a “geothermal system” for which a geothermal potential can be defined So if reinjection is considered, as in almost all the geothermal projects today, the useful temperature difference is T Tgeo Trej , being Trej T0 (see Fig. 2). This means that a mass flow rate higher than M should be extracted for power production, according to the upper limit given by

M * c p Tgeo Trej

M

(1)

(Tgeo T0 )

It is important to understand that the upper limit for the energy potential is a function of the whole “geothermal system”, as defined in the Introduction and remarked in Fig. 3

f geothermal system

(2)

The meaning of the last two equations is that, for a given value of maximum energy potential and for a given value of T, a maximum value of mass flow rate (extracted/reinjected), M*, can be defined. M* clearly depends on a lot of parameters and factors (both natural and technological): permeability distribution; hydraulic linking between the production and the reinjection areas; siting of the wells; natural recharge (meteoric water) to the reservoir. Due to the presence of this upper limit it is clear that the value of T used is inversely proportional to the value of M*. So in case of Trej increasing, induced by unacceptable chemical deposition phenomena, the mass flow rate M increases, and it could reach excessive values, causing unwanted cooling down of the whole aquifer. Consequently each reservoir presents an optimal combination of mass flow rate extractable and reinjection temperature and a correct design should follow this rule. As it can be seen, this optimization problem involves the whole “geothermal system” (plant, reservoir, environment and their mutual links), so a multidisciplinary approach become necessary. The difference between the temperature of the reservoir (Tgeo) and the reinjection temperature (Trej), together with the geofluid availability ( M geo in the following, being equivalent to M ) and the environment reference temperature (T0) contribute to define the exergy and energy potential of the geothermal field. The power production available from a plant can be defined according to the law

Wnet

I

M geo (hgeo hrej )

I

M geo c p , geo T , p (Tgeo Trej )

(3)

where I is the First Law Efficiency of the plant, that is a complex function of the temperature difference (Tgeo-Trej), of the condensation temperature (strongly influenced by the environmental temperature) and of the particular Organic Rankine Cycle (ORC) used in the binary plant. Considering the specific enthalpy of the geofluid I can be defined as follows Wnet (4) I M geo (hgeo hrej ) 42

2.2 Binary power plant design and performance parameters An important performance parameter for the analysis of the geothermal plant is the mass flow rate to generate a fixed power output, or specific brine consumption, which is given by: M = geo (5) Wnet It is inversely proportional to the First and Second Law Efficiency, which are defined as Wnet (6) I M geo hgeo hrej II

Wnet M geo egeo

Wnet

(7)

M geo hgeo T0 sgeo

Previous studies, [3-4] show that the specific brine consumption strongly depends on the difference between reservoir temperature (Tgeo) and reinjection temperature (Trej), varying from 25-40 kg/s for each MW produced in case of source temperature of 150-160 °C, up to over 100 kg/s for each MW produced in case of Tgeo = 110 °C. Due to the medium-low specific enthalpy entering the heat exchanger, this typical values of flow rate have to be considered in terms of design parameters and costs, particularly for small size power plants. The output power depends mainly by the thermodynamic cycle typology of the secondary fluid. Considering the schemes of Fig. 4 and 5 it is Wnet

Wgross W pump WCS

mw

fluid

hesp

vw

fluid

psat

pcond

(8)

WCS

pump

mw

h3 h1

fluid

M geo hgeo hrej

(9)

A limited reduction of the temperature of the geothermal source from Tgeo to T*< Tgeo, can be compensated by an increase of the mass flow rate extracted. This is possible if the balance

M * h* hrej

M geo hgeo hrej

(10)

can be maintained even if there is a decline of temperature and pressure in the reservoir, because

h*

hrej

c T * Trej

(11)

(a)

(b)

Fig. 4. Thermodynamic cycles used in binary power plants (fluids R600a and R134a) A temperature reduction of the geothermal source could cause the end of life of the plant because it could be impossible to maintain a correct pinch-point value in the heat exchanger (in Fig. 5 the decrease of the rejection temperature profile, caused by a decrease of the source temperature, cause a decrease of pinch point from the value PP1 to the value PP2). This problem is important for each typology of ORC, but in particular with reference to advanced heat recovery solutions, like Rankine with superheater, Kalina and Supercritical cycles and for Tgeo < 120-130 °C. 43

T T geo Geofluid temperature profile

PP1

PP2

3

2 T rej

Working fluid profile

1

h

Fig. 5. Temperature profiles trend in the heat exchanger of a binary plant

3. Design strategy for binary cycle power plant The methodology proposed has the purpose of defining peculiar targets which are characteristic of small size ORC technology. Geothermal resource is totally renewable only under particular conditions, that have to be assured by a sustainable conception of a geothermal utilization project. One of the main aspects remarked in this paper is that the evaluation of the effective sustainability and durability of a project is possible only under the appropriate values of (considered as the main communication parameter between plant and reservoir) and of the plant size itself. This approach can be easily extended to thermal uses and medium size power plants. As it has been previously discussed for the geothermal potential assessment, also for the reinjection strategy a general approach can't be defined for every kind of geothermal field. However general validity targets can be individuated, with the main task of maximum sustainability and useful lifetime of the plant. The mutual siting of the production and reinjection wells is often the result of a tough decision process. Numerical simulation of the reservoir circulation system is surely a useful instrument in this operation. The reinjection strategy has the role of keeping the design nominal conditions of extraction to be met for a long lifetime period. Off-design operation of such plants can be very penalizing for the efficiencies and fluid rate consumption (reservoir impoverishment). Each coupling between a utilization plant and a particular geothermal resource would have a single optimized thermodynamic cycle and T. The coupling itself is going to be considered an optimization parameter, this means that is not given as a prescribed value. Through a numerical model, which output is the evolution and consequent response of the reservoir under a certain exploitation condition, it would be possible to globally implement this methodology. The optimization of the global “geothermal system”, above described, is the synthesis of this approach. Numerical simulations can be run for forecast of future utilization or to reproduce past histories of field utilization. In forecast mode, the specific enthalpy of the geofluid extracted vs. time can be related to a model of the plant power output. Efficiency and energy production can then be estimated from this simulated parameters. A constraint to the off-design condition can be considered, for example to a minimum acceptable percentage of the nominal power output (80 % of Wnet) to guarantee a fixed lifetime (30 years) of the plant as a possible target. In the case of fixed design power output an increasing fluid rate should be withdrawn, if T * Tgeo drops, causing a decrease of the lifetime and resource durability. Connecting the numerical simulation results (temperature distribution and fluid circulation) to the plant parameters, if the working configuration is requested to be close to the nominal condition, the profiles of specific enthalpy of the geofluid extracted and power output (vs. time) should be similar. It is evident that the elaboration of a reliable numerical model of the reservoir has a great importance for the sustainability of a geothermal project. The objective of this study is to propose a methodology for the design of a binary plant using a quite low source temperature. The methodology, of general validity, has been applied to the case of a particular area in Southern Tuscany (near Monterotondo Marittimo, Grosseto, Italy), where a geothermal resource has been 44

estimated to be available at 400-500 m below the ground level. The study on this field is still ongoing, by other researchers collaborating with the authors, and a detailed description will be object of a further paper. Let us define the resource by the following data (qualitatively): Tgeo = 110 - 120 °C; p = 15 bar and Trej = 70 °C. A condensation temperature range Tcond = 30 - 35 °C is considered. The design of the plant is carried out considering the values of used in some existing small size plants, in geothermal areas with similar resources (Table 3). Table 3. Plants tested for adaptation at the geothermal field under analysis

Bad Blumau (Aus) Wineagle (USA) Neustadt (Ger) Simbach (Ger)

%

%

kg/kJ

W kW

1,89 2,41 1,26 0,96

24,57 24,66 15,12 15,11

0,1171 0,0899 0,1916 0,3125

250 300 230 200

Tsat

Tcond

Tgeo

Trej

°C 86,5 66 80 65,8

°C 30 30 30 30

°C 110 110 98 78

°C 86 70 78 65

3.1. Numerical simulation of the geothermal reservoir The models of the geothermal reservoirs are of unquestionably importance even if some limitations and criticalities are well known [3], [5] and [16]. The results reliability strongly depends from the accuracy level of the input data (thermophysical parameters, initial conditions and boundary conditions). Usually these data are known with different precision during the steps of the geothermal project, so a hard work of refinement is necessary to adapt the model to the progressive work of exploration. A schematic workflow for the realization of a numerical model is provided in Fig. 6. Calibration is an important step for the model definition. The possibility of “inverse modeling” approach is well known. It could be important to start with simple models (also lumped parameter models) to clarify the conceptual scheme and the physical consistence of the problem.

Fig. 6. Conceptual workflow for the realization of a numerical model of a geothermal reservoir The importance of fields like the one considered is remarkable. In Italy (and generally in the proximities of the main high enthalpy fields) this kind of medium temperature reservoir are going to be exploited in the next few years by a lot of industrial subjects. A numerical model of the Monterotondo Marittimo area of Larderello field (Italy) has been realized using the commercial software Petrasim (in which is implemented the TOUGH2 simulator), [17]. The model domain extension and the various materials used are shown in Fig. 7. The conceptual model of the field is not an aim of this paper, its development is still ongoing in collaboration with other researchers. It will be covered by a further paper the authors are involved in. The model presented can be considered to be a good qualitative representation of the reservoir, and it is here used to elaborate 45

and underline some specific features of the sustainable design methodology of an ORC power plant. Sensitivity analysis and extension of the scale of the domain are future developments of this model.

COVER RESERVOIR FORMATIONS HOT FLUID RECHARGE BASEMENT

Fig. 7. Model domain and sketch of the main structural features of a numerical model grid of a geothermal reservoir The model has been built in order to elaborate the optimal production/reinjection strategy for an ORC installation. Simulations of the natural steady-state (unperturbed) conditions of the field have been firstly run and then different exploitation scenarios have been simulated.

3.1.1. Simulation of the exploitation scenarios (ORC power plant) The study of the particular geological structures and other exploration features will be objects of a different paper in which the authors are involved. Production scenarios have been realized to study the reservoir response to exploitation conditions and to design a possible industrial utilization of the field. The exploitation scenarios simulated are relative to production/reinjection of fluid in case of the presence of an ORC power plant with the following characteristics: size of the plant: 200 - 1000 kW (mass flow rate 15 - 100 kg/s) reinjection temperature fixed at 70 °C Model results will be now discussed, linked to the values and geothermal fluid mass flow rates extraction/reinjection corresponding to relative extraction temperature decrease with time and fixed power output. The value of 15 kg/s the mass flow rate is an average value for the operation of a plant like those described in Table 3, for a plant size of about 200 kW. With this low value of mass flow rate a complete sustainability of the plant is possible, because temperature reductions of 2 °C in 30 years and of about 4 °C in 50 years (Fig. 8). 95 93 91

T [°C]

89 87 85 83 15 kg/s extraction rate

81

50 kg/s extraction rate 79

100 kg/s extraction rate

77 0

5

10

15

20

25

30

35

40

45

50

Time [years]

Fig. 8. Simulation of the production scenarios: temperatures at the production well (extraction/reinjection mass flow rates: 15 kg/s, 50 kg/s and 100 kg/s). A mass flow rate extraction of 50 kg/s (for a power production of about 500 kW) determines a temperature decrease of the source of about 6 °C in 30 years and about 10 °C in 50 years, this 46

would be critical for the plant so a sufficient life of the plant is not assured. Besides the extraction of a mass flow rate of 100 kg/s (that would permit a power output of about 1 MW) could appear to be unsustainable for this geothermal field. The diagram of temperature reduction during the lifetime of the plant is provided in Fig. 8.

Fig. 9. Temperature iso-surfaces in the scenario with two production wells and one reinjection well. 93

107

15 + 15 kg/s extraction rate 50 + 50 kg/s extraction rate

91

15 + 15 kg/s extraction rate 50 + 50 kg/s extraction rate

102

89 97

87

92

T [°C]

85 83

87

81 82

79 77

77 0

a)

10

20

30

40

50

0

b)

Time [years]

10

20

30

40

50

Time [years]

Fig. 10. Temperature evolution for the “PROD1” well (a) and for the “PROD2” well (b). It is possible to observe temperature decreases of about 11 °C after 30 years of exploitation and 15 °C in 50 years. In both the last two cases it would be difficult to maintain a correct working of the ORC plant. A further layout scenario in which two production wells can be considered (“PROD1” and “PROD2”, in Fig. 10) and one reinjection well (where the sum of the extracted flow rates is reinjected). In Figs. 9-10 the simulated extraction temperatures evolution is shown (for a period up to 50 years). For both the production wells the extraction rate of 50 kg/s is unsustainable.

4. Discussion and conclusions The success of the application of the binary plants depends on a correct design strategy. Their efficiencies are extremely sensitive to external parameters (available T of fluid, environment temperature, fluid pressure, permeability) changes. The characterization of the resource available is a fundamental initial step of each design process and it appears to be more important than the optimization of the plant itself (combination of working fluid and thermodynamic cycle). Sustainable geothermal power production refers to the optimal extraction/reinjection fluid rate which should be maintained for a very long time, according to certain design parameters of the plant and of the wells system. This task requires an approach that join plant engineering and reservoir engineering in order to avoid overexploitation and pursuing energy-efficient utilization. Careful monitoring and exploration are essential for sustainable reservoir management. The case 47

study here considered has a general validity value and can be extended to more complex situations, although the model has to be considered as qualitative. Numerical simulation of geothermal reservoirs is a very useful and strategic instrument to elaborate exploitation scenarios and to define a correct reinjection strategy mainly in case of moderate temperature of the source. Its reliability strongly depends on the accuracy of the input data. The analysis has been developed with a multidisciplinary approach to geothermal systems exploitation considering the interconnections between Geoscience and Energy Engineering, concerning in particular the geothermal energy assessment. The design of a small size geothermal plant has been carried out for the exploitation of a geothermal resource in the Larderello (Italy) geothermal area (Monterotondo Marittimo). A moderate temperature geothermal source (110-120 °C), estimated to be available at relatively low depth (400-500 m below the ground level) has been considered. The adaptability of plants size of about 200 kW or discrete multiples (up to 200 kW x 5 = 1 MW) is analyzed with the support of a numerical model of the reservoir in order to elaborate a production/reinjection strategy. According to the qualitative model elaborated, a plant size of 200 kW could be run sustainably for a period of almost 30 years; the geofluid rate is estimated to be not higher than 20 kg/s. Higher fluid rates (for example twice the previous size) would be critical for the resource durability. The extraction of a mass flow rate of 100 kg/s, that would permit a power production of the order of magnitude of 1 MW, appears to be unsustainable.

Acknowledgments The authors want to acknowledge Vera Trinciarelli (Bachelor Degree in Energy Engineering at the University of Pisa) and Arianna Secchiari (Master Degree in Geological Sciences at the University of Pisa) among others, for their contribution in the Monterotondo numerical model.

Nomenclature c e h p m M Q T To W v

specific heat at constant pressure, J/(kg K) specific exergy, J/kg specific enthalpy, J/kg pressure, bar mass flow rate of the working fluid, kg/s mass flow rate of the geothermal fluid, kg/s heat flow rate, W temperature, °C reference temperature, K power. W specific volume, m3/kg

Greek symbols specific brine consumption, kg/MJ specific thermal capacity of the reservoir, kJ/kg K T temperature difference, °C efficiency I First Law efficiency II Second Law efficiency function of the whole “geothermal system” Subscripts and superscripts cond at the condenser CS of the cooling system geo of the geothermal brine 48

gross liq net pump rej sat w-fluid *

gross power of the saturated liquid net value (of the power) of the pump rejection saturation of the working (organic) fluid value of mass flow rate, temperature and enthalpy of the reservoirs after exploitation

References [1] [2] [3] [4] [5] [6] [7] [8]

[9] [10] [11] [12] [13] [14] [15] [16] [17]

Barbier E., Geothermal energy technology and current status: an overview. Renewable and Sustainable Energy Reviews 2002; 6: 3–65. Stefansson V., World Geothermal Assessment. Proceedings World Geothermal Congress 2005; 2005 Apr 24-29; Antalya, Turkey. Franco A., Vaccaro M., An integrated “Reservoir-Plant” strategy for a sustainable and efficient use of geothermal resources. Energy 2012; 37: 299-310. Axelsson G., Sustainable geothermal utilization e Case histories; definitions; research issues and modelling. Geothermics 2010; 39: 283-291. Vaccaro M., Franco A., Casarosa C., Strategies for the Sustainable Use of Geothermal Reservoirs. In: Conference Proceedings of the first Sustainable Earth Sciences Conference; 2005 Nov 8-10; Valencia, Spain. Cataldi R., Muffler P., Methods for regional assessment of geothermal resources. Geothermics 1978; 7: 53-89. Hochstein, M.P., Classification and assessment of geothermal resources. In: Dickson, M.H., Fanelli, M. editors. Small Geothermal Resources. A Guide to Development and Utilization, Rome, Italy, 1990. pp. 31–57. Benderitter Y., Cormy G. Possible approach to geothermal research and relative cost estimate. In: Dickson M.H., Fanelli M., editors. Small geothermal resources. A Guide to Development and Utilization, Rome, Italy, 1990. p. 61–71. Nicholson K., Geothermal Fluids. Springer-Verlag, New York, 1993. Axelsson G., Gunnlaugsson E. Long-term Monitoring of High- and Low-enthalpy Fields under Exploitation. World Geothermal Congress, Kyushu Japan, 2000, 226 pp. Bertani R., Geothermal power generation in the world 2005–2010 update report, Geothermics 2012, 41: 1-29. Hettiarachchi H.D.M., Golubovica M.,Woreka W.M., Yasuyuki I., Optimum design criteria for an Organic Rankine cycle using low-temperature geothermal heat sources. Energy 2007; 32: 1698–1706. Franco A., Villani M., Optimal design of binary cycle power plants for water-dominated, medium-temperature geothermal fields. Geothermics 2009; 38: 379–391. Tchanche B.F., Lambrinos Gr., Frangoudakis, A. Papadakis G., Low-grade heat conversion into power using organic Rankine cycles – A review of various applications, Renewable and Sustainable Energy Reviews 2011; 15: 3963-3979. Franco A., Power production from a moderate temperature geothermal resource with regenerative Organic Rankine Cycles. Energy for Sustainable Development 2011; 15: 411419. O’Sullivan M.J., Pruess K., Lippman M.J., State of the art of geothermal reservoir simulation. Geothermics 2001; 30: 395-429. ThunderHead engineering. PetraSim 5 user manual. (http://www.thunderheadeng.com). 49

PROCEEDINGS OF ECOS 2012 - THE 25 TH INTERNATIONAL CONFERENCE ON ENERGETIC AND ENVIRONMENTAL BENEFITS FROM WASTE MANAGEMENT: EXPERIMENTAL ANALYSIS OF SUSTAINABLE LANDFILL JUNE 26-29, 2012, PERUGIA, ITALY

Energy and Environmental Benefits from Waste Management: Experimental Analysis of Sustainable Landfilling Francesco Di Mariaa, Alessandro Canovaib, Federico Valentini,b Alessio Sordia, Caterina Micalea a

Dipartimento di Ingegneria Industriale, Perugia, Italy, [email protected] b GESENU spa, Perugia, Italy, [email protected]

Abstract: Sustainable landfilling is an innovative way of managing waste disposal plants, leading both to environmental and energy benefits. The rapidity of the anaerobic biodegradation process of the biodegradable fraction of urban waste can be improved by combining aerobic with anaerobic treatment. Aerobic pre-treatment of the waste, ranging from 15 to 60 days, together with leachate recirculation, can lead to the production of more 3 than 170 Nm per tonne of Volatile Solids disposed of in the landfill. The anaerobic biodegradation process can take from 3 to 14 years, depending on some differences in landfill management. Further, the total amount of renewable electrical energy able to be recovered ranges from 42 to 44 GWh, and the biomethane loss can be reduced to less than 39% of the total production.

Keywords: Anaerobic biodegradation, Biomethane, Bioreactor, Landfill Gas, Renewable energy

1. Introduction In the past, sanitary landfills or dumps were often the only way that Municipal Solid Waste (MSW) was managed around the world [1], [2]. As known [3], under anaerobic conditions, MSW can spontaneously generate a large amount of biological gas with a high content of methane. This phenomenon occurs when the MSW is disposed of in landfills or dumps. The management of MSW is one of the most relevant sources of anthropogenic Green House Gas (GHG) [3], from about 3 to 13% of the total amount. In the EU27 [4], the total production of MSW is about 260 Mtonnes (2008) and about 105 Mtonnes (42%) are managed directly by landfilling. Even if the amount of landfilled waste has been reduced significantly in the last years, it is, nevertheless, a current serious environmental hazard and an important source of renewable energy. The theoretical amount of Landfill Gas (LFG) able to be produced per tonne of landfilled MSW has been estimated to be about 130-160 Nm3 [5], with a methane fraction ranging from 45 to 55 %v/v [6]. The remaining fraction is mainly carbon dioxide (from 55 to 40 %) together with other macro and micro concentration compounds [3], [6]. These values can be significantly influenced by many factors (i.e. waste composition, collection, pre-treatment, landfill management,…), but in any case they show the importance of the phenomenon and the need to provide suitable management to avoid risks to human health and the environment. Furthermore, the LFG is produced from the biological degradation of biodegradable waste and for this reason is considered a renewable energy source. The amount of renewable energy produced from LFG [7] in the EU25 was about 3,000 ktoe in 2010, with about 12% coming form Italy. Unfortunately, as a consequence of the most commonly used landfill management techniques, also consisting in achieving maximum separation between the 50

disposed waste and the external environment, a large fraction of the potential energy of LFG is missed. This is due to the loss of LFG during waste disposal activities, before and during energy recovery system implementation, and to typical inhibition phenomenon occurring once the anaerobic process starts, which can affect the methanogens phase. In fact, this last effect causes a large fraction of LFG to be produced at such a low rate for a long period of time that energy recovery is not sustainable. Furthermore, the low degradation rate of the organic material in the waste also causes a high pollutant content of the landfill leachate for a long time, making its management a serious and costly problem. All these negative aspects can be dealt with using a different management concept, based on the sustainable landfill scheme [8]. The aim of this alternative management strategy is to enhance the natural, biological process occurring in the body of the landfill in order to achieve a more rapid stabilization of the disposed waste. According to this concept, the landfill is managed as a bioreactor. This can lead to some advantages such as an increase in energy recovery from the LFG along with reduction of the time required to reach a high level of stabilization of the waste. Assuming a constant LFG potential of the waste disposed of in the landfill, theoretically a bioreactor landfill should be able to produce the same amount of gas in less time (Fig. 1). This would reduce the LFG losses, leading to further increase in recoverable energy, due to the exploitation of a larger size Internal Combustion Engine (ICE) with a greater electrical efficiency [9]. One of the methods used for enhancing biodegradation processes is recirculation of the leachate, to provide an adequate moisture content and bacterial distribution [10], [11] [12]. The enhanced bioconversion of the waste into LFG also produces an increase in the waste settlement process, leading to a further advantage, which is the increased amount of waste actually able to be disposed of in the landfill per unit of volume. Also the amount of polluting leachate can be significantly reduced. Appropriate treatment of the waste before being disposed of can improve the biological stabilization process pursued by the sustainable/bioreactor landfill management along with a reduction of LFG loss. In this study, the effect of mechanical and aerobic pre-treatment of the LFG produced from waste disposed of in a bioreactor landfill was analyzed. The study focused on the MSW collected and managed in a given Italian area, and the possible energy and environmental benefits were also been evaluated.

Fig. 1. Comparison of LFG production by traditional and bioreactor systems .

2. System description and Methods 2.1. The MSW management area

51

The MSW management system considered was Management District (MD) n°2 of the Umbria Region in central Italy (Fig. 2). The amount of waste produced in this area is about 250,000 tonnes per year and the resident population is about 350,000 inhabitants (Table 1). In MD n°2 only one Mechanical Biological Treatment (MBT) plant operates, treating the NonDifferentiated Waste (NDW) resulting after source selected collection. The aim of the MBT facility is to increase the amount of recovery of waste materials and to reduce both the mass and the reactivity of the residual waste mass before landfilling. This aim is pursued by mechanically sorting the NDW, by metal separation and size screening, followed by biological treatment. The screening is basically done using trommels with 100mm diameter holes, which produce an oversize stream of dry components (i.e. plastic, paper and textile), with a high Lower Heating Value (LHV) and is exploitable as fuel in co-combustion or in Waste-to-Energy (WtE) plants. The undersize stream, representing about 50% w/w of the NDW entering the plant, is rapidly biodegradable materials (i.e. yard trimming, household waste, kitchen residues) and is further treated in the aerobic biological section of the MBT plant. In this section aerobic biodegradation of the material stabilizes the waste along with a significant mass reduction. After this treatment the undersize material is disposed of in the existing landfill, which is close to the maximum disposal volume allowed. At the beginning of 2012, the construction of a new landfill was started, which includes leachate recirculation. Table 1. Main features of MD n°2. Parameter MSW (2016) Inhabitants NDW at MBT inlet Undersize (1)

Value 255,575 349,703 476 50

Unit Tonnes/year Tonnes/day % of NDW(1)

Evaluated at trommel outlet.

Fig. 2. Umbria region and position of Management District n° 2 ( MDn°2). 52

Due to the large amount of the undersize and its rapid biodegradability, a large amount of LFG can be produced once it is disposed of in the bioreactor landfill. Unfortunately, due to the high Total Solids (TS) content and low pH value of the undersize, some inhibition phenomenon can occur during anaerobic degradation. To avoid or manage this, leachate recirculation and aerobic pretreatment before disposal could be solutions. This last aspect is the focus of the present analysis.

2.2. Waste characterization Waste samples were characterized both by make up of the components and by chemical and physical analyses. The components were analyzed manually as Non-Biodegradable (NB), Biodegradable and fines (20S) (i.e. particles [accessed 16.11.2011]. [32] Solaronix SA, Switzerland – Available at: http://www.solaronix.ch> [accessed 14.11.2011]. [33] G24 Innovations. Wales (UK) – Available at:http://www.g24i.com> [accessed 15.11.2011]. [34] Keoleian G.A., Lewis G.M ., Application of life cycle energy analysis to photovoltaic module design. Progress in Photovoltaics: Research and Applications 1997;5:287-300. [35] Pearsal N., Science, Technology and Applications Group of the EU Photovoltaic Technology Platform, A Strategic Research A genda for Photovoltaic Solar Energy Technology, Photovoltaic Technology Platform Edition 2. Newcastle upon Tyne, UK: School of CEIS, Northumbria Photovoltaics Applications Centre. 2011 Sep. ISBN 978-92-79-20172-1.

132

PROCEEDINGS OF ECOS 2012 - THE 25 TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

Low CO2 emission hybrid Solar CC power system Yuanyuan Li a, Na Zhang b, Ruixian Cai c a

c

Institute of Engineering Thermophysics, Chinese Academy of Sciences; Graduate University of the Chinese Academy of Sciences, P.O. Box 2706, Beijing, P.R. China, [email protected] b Institute of Engineering Thermophysics, CAS, P.O. Box 2706, Beijing, P.R. China, [email protected] (CA)

Institute of Engineering Thermophysics, CAS, P.O. Box 2706, Beijing, P.R. China, [email protected]

Abstract: Based on the principle of cascade utilization of multiple energy resources, a novel concept for gas-steam combined cycle integrated with solar thermo-chemical conversion and CO 2 capture, named low CO 2 emission hybrid Solar CC power plant (LEHSOLCC), has been proposed and analysed. The hybrid power system uses methane as its input fuel. The collected solar heat at 550 oC is applied to provide heat for the endothermic methane reformation. The reforming reaction is integrated with a hydrogen separation membrane, which continuously withdraws hydrogen from the reaction zone and enables the chemical equilibrium to shift towards the product side. The pure H2, collected in permeate side, fuel a topping Brayton cycle, and the exhaust drives a triple-pressure reheat Rankine bottoming cycle to produce additional power. The produced syngas in retentate zone is enriched with CO2 (81.8%v) and thus can be suitable to be processed with precombustion decarbonization. In the proposed power system, the low level solar heat is first converted to syngas chemical exergy via reforming, and then released as high-temperature thermal energy in an advanced combined cycle system for power generation, thus achieving its high-efficiency heatpower conversion. To reduce the exergy destruction, special attention is paid to the thermal match of the internal heat recuperation, as well as to the thermo-chemical match between the solar heat and the reforming process. The system is thermodynamically simulated using the ASPEN PLUS code. The results show that with 91% CO2 captured, the specific CO2 emission is 25 g/kWh. Exergy efficiency of 58% and thermal efficiency of 51.6% can be obtained. CO2 capture brings about 8.4%-points thermal efficiency penalty compared with a gas-steam combined cycle system at the same technical level without CO 2 capture, but exergy efficiency remains the same level as the reference system. Fossil fuel saving ratio of 31.2% is achievable with a solar thermal share of 28.2%, and the net solar-to-electric efficiency, based on the gross solar heat incident on the collector, is about 36.4% compared with the same gas-steam combined cycle system with equivalent CO 2 removal rate by way of post-combustion decarbonization.

Keywords: Hybrid Power System, Solar Thermal Energy, Membrane Reformer, Thermo-chemical Conversion, CO2 capture.

1. Introduction Solar thermal conversion is a promising technology for power generation. It is an efficient means to reduce emissions and to save fossil fuel. Generally speaking, higher working fluid temperature is preferable in power system for achieving higher heat-to-power efficiency. At low ones (~200oC or below) solar-only thermal power generation has low efficiency. For instance, the efficiency of Rankine cycle systems using organic working fluids is generally lower than 10% [1, 2]. Solar heat collected at high temperatures, however, associates with a significant increase in solar plant costs and reduced collector efficiency. In order to minish the thermo-economic gap between conventional

Corresponding author: Tel.: 86-10-82543030; fax: 86-10-82543019. E-mail address: [email protected].

133

power plant technologies and solar plants, it is necessary to reduce the cost of the solar specific components and to improve the solar heat-to-electricity conversion efficiency. Solar hybrid systems, in which solar heat and fossil fuel are used in a complementary way, provide a attractive solution for achieving high efficiency converstion of solar heat. Solar heat can be integrated into the power system either thermally or thermo-chemically. The later comprises solar upgrading of hydrocarbons by some endothermic reactions, and using the upgraded hydrogen-rich fuel to generate power in high efficiency conversion systems such as gas turbines and fuel cells. Methanol-steam reforming and methanol decomposition can achieve over 90% conversion into H2 -rich syngas at around 250°C. By taking advantage of the high conversion rates at this relatively low temperatures, Jin and co-workers proposed a combined cycle (called Solar CC) that ingeniously integrates low/mid-temperature solar thermal energy with methanol decomposition [3]. By heating the endothermic decomposition reaction, solar thermal energy is upgraded to chemical energy of the produced syngas. In a case study [3], exergy efficiency of the hybrid combined cycle system is 60.7%. The net solar-to-electricity efficiency can reach 35% with the solar thermal share of 18%, and the CO2 emission is 310 g/kWh without regard of CO 2 capture. For natural-gas fired power plants, the methane-steam reforming generally requires above 800oC with Ni-based catalyst to obtain high methane conversion. Tamme presented a high temperature (>1000oC) solar hybrid system comprising solar upgrading of methane by steam reforming in solar specific receiver-reactors and utilizing the upgraded H 2 -rich fuel in advanced gas-steam combined cycles [4]. In comparison to a conventional CC system, about 30% of fuel can be saved. However, due to the high temperature solar heat collection, the upgrading in energy level of solar thermal energy to syngas chemical energy is limited and the cost of the solar components is considerably high. To avoid the high cost and low collecting efficiency caused by high temperature solar heat collection, moreover, to allow the low temperature solar heat to achieve its high-efficiency heatpower conversion, Zhang and Lior proposed a solar-assisted chemically recuperated gas turbine system (SOLRGT) with indirect solar heat upgrading [5]. Solar heat collected at ~220°C is used to generate steam for methane reformation, thus first transformed into vapour latent heat, and then converted to the produced syngas chemical energy via the reforming reaction. The upgraded solar fuel is eventually burned in a high-efficiency power system. About 20-30% fossil fuel saving ratio can be achieved compared with a conventional chemically recuperated gas turbine system (CRGT) without solar heat contribution. However, nearly 80% of the total energy input is provided by methane, leading to 342.7g/kWh CO2 emission. In the SOLRGT system[5,6], the reforming section retrieves heat from the turbine exhausts at a temperature below 600oC, resulting to a limited CH4 conversion of 37.8%, far below the desired 95% conversion rate for pre-combustion decarbonization employment. Higher reforming temperature elevates methane conversion. Boosting the reaction temperature by means of supplementary firing with additional fossil fuel, however, imposes penalty on the overall system efficiency. Zhang et al proposed a zero CO2 emission SOLRGT system [7,8] based on oxy-fuel combustion. Besides raising reforming temperature, high CH4 conversion can be achieved also with selective removal of reaction products, such as H2 or CO2 . Palladium membrane, whose mechanism is based on a H 2 solution-diffusion mechanism on perm-selective film, has been applied to construct membrane reactors to shift thermodynamic equilibrium limited reactions. The investigation carried out has showed that 99% CH4 conversion can be obtained at 550oC and 3~9bar [9]. Provided that integrating the Pd membrane reactor with the SOLRGT system, the pre-combustion decarbonization may be feasible. Based on this consideration, in the present paper, we propose a low CO2 emission hybrid Solar CC cycle (LEHSOLCC), in which solar heat collected at middle temperature (~550oC) is used to heat the endothermic steam reforming of methane in a membrane reactor, achieving nearly full CH4 conversion, realizing not only the high-efficiency cascade utilization of fuel chemical exergy and solar thermal energy, but the CO2 capture prior to combustion with low energy 134

penalty. A design-point performance analysis shows that the system attains a net exergy efficiency of 58% and specific CO2 emissions of 25 g/kWh with 91% CO2 capture rate. Fossil fuel saving ratio of 31.2% is achievable with a solar thermal share of about 28.2%. The net solar-to-electric efficiency, based on the gross solar heat incident on the collector, is about 36.4% compared with a gas-steam combined cycle system with post-combustion decarbonization at the same fossil fuel input.

2. System configuration description 2.1. Reforming in membrane reator Figure 1 shows the schematic diagram of methane reforming in a palladium membrane reactor. The Pd membrane is prepared via a novel electroless plating method [10]. The alumina support includes glaze and porous part to plating. One end of the support tube is sealed. The open end is applied for the introduction of a sweep gas flux and collection of permeated H 2 , simultaneously. The Pd membrane tube is located inside the center of a dense stainless steel tube reactor. An annulus chamber between the stainless steel tube and membrane tube is thus applied for reaction, which also served as the retentate side of the Pd membrane. While the inner volume of the membrane tube is the permeate zone. Diluted with silica, the catalyst is packed to construct a catalyst bed, extended over the membrane. The reactor is firstly heated to reaction temperature of 550oC, after stabilizing, the mixture of methane and steam is gradually introduced to the reactor. As the reforming processes, reaction products H2 and CO2 in reaction zone concentrate and their pressure keep increasing. Driven by the pressure difference, H2 moves to the permeate zone continuously, and is taken away by the sweep gas; and the impermeable CO 2 gas is collected from the bottom of the reaction zone. Note that the sweep gas flow should be countercurrent to the feed stream in order to maintain a high H 2 partial pressure drop across membrane to promote H 2 permeation.

Sweep gas Sweep gas+H2

CH4+H2O Pd membrane Catalyst Reactor

Solar heater

Pd membrane tube

CO2-rich gas Fig. 1. Schematic diagram of membrane reformer Figure 2 shows the principle of the membrane reformer [11]. The driving force of the membrane transport is the difference in H2 partial pressures between the retentate and the permeate sides. The molar flow rate F of H2 through the membrane may be described by the transport equation: F=B·A· [(PH)n-(PL) n] [11], where B is the H2 permeance efficiency, A is the membrane surface, and (PH)n -(PL)n is the partial pressure difference across the membrane. It can be seen that separation performance of membrane depends on two factors, flux and selectivity. Flux is determined by knowing the mass of permeate collected, membrane area and the running time. H 2 selectivity represents the measure of the preferential transport of H 2. It is defined by the H2/N 2 separation coefficient of membrane, which is determined with pure H 2 and N2 at constant pressure drop and 135

temperature across the membrane. It can be found that no permeation of nitrogen across the Pd membrane is detected and an almost infinite H2 selectivity can be obtained [9]. Moreover, the Pd membrane is experimentally characterized by gas permeation tests at different temperature and pressure using pure gases (H2, N2, CO, CO2 , CH4 and H2O). It is observed that only hydrogen permeates through the membrane [12], suggesting that Pd membrane has high H 2 permeation and separation performance. Additionally, the selectivity to a carbon product (CO or CO 2 ) by current reaction is defined as the ratio of molar flow rate of the selected product to the converted methane molar flow rate. To increase the CO2 capture rate, the CO selectivity should be suppressed as far as possible. Qsol CO+3H2 CH4+H2O CO+H2O CO2+H2 H2 H2 Reaction side H2 Permeate side n n F=BA(PH -PL ) Sweep gas+H2

CO2-rich gas

CH4+H2O

Pd membrane Sweep gas

Fig. 2. Principle of the membrane reformer

2.2. Low CO2 emission Hybrid Solar CC system A gas-steam combined cycle integrated with solar thermo-chemical process and CO 2 capture, named LEHSOLCC (Fig. 3), has been set up to demonstrate the solar-driven-membrane reaction in power plant. A middle temperature (550oC) tower-type solar collector is introduced to provide heat for the endothermic methane reforming at pressure of 9 bar and steam-to-carbon ratio of 3.5. The reformer is integrated with a hydrogen separation membrane, enabling continuously removal of hydrogen from the retentate (reaction) zone, and thus shifting the reaction to the product side. The pure H 2 (13), collected in permeate side, warms up the reforming reactants (7, 10) and fuels a topping Brayton cycle after being compressed. The CO2 -rich syngas (12) in retentate zone is first used to preheat the reforming reactants (7, 10), then cooled down to near ambient temperature and processed in a CO2 physical absorber, where more than 90% CO2 is removed. The remainder (22), which mainly contains 68.8%vH 2, 9%vCO, 3.8%vCH4 and 18.4%vCO2, is compressed, and also utilized as the fuel in the topping Brayton cycle. The gas turbine exhaust (27) eventually drives a triple-pressure reheat Rankine bottoming cycle to produce additional power. Together with the complementary make-up water (5), the condensate (18) drained from the CO2 stream is recycled to the reformer. To reduce the exergy destruction, special attention is paid to the thermal match of the internal heat recuperation, as well as to the thermo-chemical match between the solar heat and the reforming process. In the proposed power system, there are two stages of solar heat collection. The low level (lowtemperature) solar heat (200oC) evaporates the reforming water and is converted into steam latent heat, and the middle level (middle-temperature) solar heat (550oC) is used to drive the endothermic reaction and transformed into syngas chemical exergy via reforming, and finally released as hightemperature thermal energy through combustion in the advanced gas-steam combined cycle system for power generation, achieving a high-efficiency heat-power conversion. With the assistance of solar heat, the temperature match in the syngas heat recuperation process can be improved since it provides only sensible heat to the reforming reactants, leading to reduced heat transfer exergy loss. Moreover, the turbine exhaust heat is available to drive a Rankine bottoming cycle to produce additional power. 136

The integration of a H2 separating membrane in the reformer achieves nearly full CH4 conversion at middle temperature. Meanwhile, the CO 2 concentration is increased and CO2 capture energy consumption is reduced. The system accomplishes a high-efficiency solar heat-power conversion and low-penalty pre-combustion decarbonization. Solar tower Air 1 Compressor

Turbine Steam turbine

Membrane reformer

2

12 11

25 CH4 Compressor 3 4

10

Superheater

28

15

17

21

16 18

Economizer

Condenser

30

CO2 physical absorber 23 CO2

Pump

Trough-solar collector 8

31

29

Pump Condenser

H2 compressor

Evaporator

27

Fuel compressor 22 24

20

14

9

32

26 Combustor

HRSG

13

7

19 Pump 6

H2O 5

Air/gas/CH4

H2O

Fig. 3. Schematic diagram of LEHSOLCC system

3. Computation model and its validation 3.1. Main assumptions for the simulation The cycles presented in the paper are modeled in ASPEN PLUS process simulation software [13]. The component models are based on the energy balance, mass balance, and species balance, with a default relative convergence error tolerance of 0.01%, which is the specified tolerance for all tear convergence variables. The RK-SOAVE and STEAM-TA thermodynamic models are selected for the thermal property calculations. The membrane reactor is modeled with a combination of Gibbs Reactor and Component Separator available in the ASPEN PLUS model library. The Gibbs Reactor determines the equilibrium conditions by minimizing Gibbs free energy, and the Separator specifies flow split fractions. The hydrogen permeation is a complex function of the partial pressure difference across the membrane and is therefore modeled with a Fortran subroutine. By assuming a certain permeability, membrane thickness and membrane area, the permeation rate can be calculated with an iterative procedure. This approach does not capture the change in flux along the length of the membrane tube, so the average flux is therefore approximated by the logarithmic mean of the flux at the inlet and outlet of the reactor [14], then it determines methane conversion and hydrogen recovery. In the solar block, the low temperature solar field is assumed to be installed with parabolic trough direct steam generation collector (DSG) [15], and the middle temperature one is with solar tower [15]. The CO2 physical absorption in the hybrid system is based on a model developed by Lozza and Chiesa [16] using the Selexol [17] absorption medium. A conventional gas-steam combine cycle system with post-combustion decarbonization is also simulated for the purpose of performance comparison, in which CO 2 capture is accomplished using a chemical absorption process (with monoethanolamine (MEA) [18] as the absorbent). Steam is extracted from the steam turbine for the absorbent regeneration, the corresponding energy and steam demands are calculated based on the given composition of the process gas. The most relevant assumptions are summarized in Table 1.

137

Table 1. Main assumptions for the simulation and calculation Parameters Gas turbine Combustor Heat exchanger Steam turbine

HRSG

Pump Parabolic trough solar collector Solar tower

Membrane reformer

Physical absorption

Chemical absorption

Inlet pressure Isentropic efficiency Pressure drop (of inlet pressure) Minimum temperature difference Pressure loss HP steam pressure RH/IP steam pressure LP steam pressure Condensing pressure HP/RH steam temperature Pinch-point temperature difference Cold side pressure drop Minimum stack temperature Efficiency Solar collector temperature Solar collector efficiency Minimal temperature difference Solar collector temperature Solar collector efficiency Minimal temperature difference Direct solar radiation Reaction Pressure Pressure drop Reaction temperature Steam-to-carbon ratio Membrane thickness CO2-to-SELEXOL mole ratio in absorbent Number of flash chambers Last chamber pressure

Number of intercoolers for CO2 compressor CO2-to-MEA mole ratio in absorbent Minimum temperature difference at solution regenerator Stripping pressure Steam supply Temperature difference in reboiler Max. gas pressure drop in absorber/stripper Number of intercoolers for CO2 compressor

Value o

Source

1300 C 88% 3% 15oC 3% 111bar 27bar 4bar 0.06bar 560oC 15oC 5% 85oC 85% ~200°C 62% 20oC ~550oC 65% 20oC 940 W/m2 9bar 5% 550oC 3.5 m 0.1 4 to obtain 95% CO2 removal 3 0.15 10°C

GT Word 2010 [20] GT Word 2010 [20] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Ertesvåg et al, 2005 [21] Zhang et al, 2012 [5] Zhang et al, 2012 [5] Zhang et al, 2012 [5] Solúcar et al 2006 [16] Solúcar et al 2006 [16] Solúcar et al 2006 [16] Solúcar et al 2006 [16] Chen et al, 2008 [9] Chen et al, 2008 [9] Chen et al, 2008 [9] Chen et al, 2008 [9] Chen et al, 2008 [9] Lozza et al, 2002 [17] Lozza et al, 2002 [17] Lozza et al, 2002 [17]

Lozza et al, 2002 [17] Lozza et al, 2002 [17] Lozza et al, 2002 [17]

1.01bar 3 bar 5°C

Lozza et al, 2002 [17] Lozza et al, 2002 [17] Lozza et al, 2002 [17]

4/10 kPa 2

Lozza et al, 2002 [17] Lozza et al, 2002 [17]

3.2. Validation of palladium membrane reactor The performance of methane reforming in Pd membrane reactor mainly depends on the membrane structure and material, catalyst and working conditions (reaction temperature, reaction pressure, 138

steam-to-carbon molar ratio, sweep gas flux, inlet gas velocity, etc.). The Pd membrane and nickelbased catalyst are characterized by high H2 permeance and fast kinetics, respectively. We consider in this paper only the influence of reaction conditions on the reaction performance. Figure 4 illustrates the influences of reaction temperature, pressure and steam-to-carbon ratio on methane conversion and CO selectivity at fixed gas velocity. Experimental data [9] and simulation result are plotted by dashed lines and solid lines with symbol, respectively, at the same working conditions. It can be seen that the simulation data and experiment results are in good agreement. Due to the instabilities of the experimental conditions, the experimental data of CH 4 conversion is slightly lower, and CO selectivity is higher than their corresponding simulation results, but all the relative differences are within 3%. Experiment Simulation

100

10 8

60

6

P=9bar S/C=3.5 I=2.6

40

4

20

2

0 440

460

480

500

520

CO selectivity (%)

CH4 conversion (% )

80

0 560

540

Reaction temperature ( C)

(a) Experiment Simulation

100

10

CH4 conversion (%)

60

6

t=550 C S/C=3.5 I=2.6

40

4 2

20

CO selectivity (%)

8

80

0

0 3

4

5

6

7

8

9

Reaction pressure (bar)

(b) Experiment Simulation

10

CH4 conversion (%)

80

8

60

P=9bar t=550 C I=2.6

40 20 0 2.4

6 4 2

2.6

2.8

3.0

3.2

3.4

CO selectivity (%)

100

0 3.6

Steam-to-carbon molar ratio

(c) Fig. 4. Effects of working conditions on reaction performance: a) Reaction temperature, b) Reaction pressure, c) Steam-to-carbon molar ratio 139

Limited by both the reaction kinemics and membrane high temperature properties, membrane reaction is normally performed in the 450~600oC range. As shown in Fig. 4, CH4 conversion monotonically increases along with temperature in the calculated range. High temperature favors methane conversion and H2 permeation, leading to high separation efficiency. Meanwhile, the behavior of H2 permeation pulls the reaction towards the product direction, offsets the negative effect of high temperature on the shift reaction, and thus efficiently suppresses CO formation. A non-monotonic variation of CO selectivity is observed. CO selectivity first increases with reaction temperature, and then slightly decreases. High reaction pressure is disadvantageous to CH 4 conversion in terms of thermodynamic equilibrium, but it results in a higher H2 partial pressure in the reaction zone, and subsequently a higher driving force for H2 permeation. Therefore, the influence of pressure is a compromise between the these two effects. It can be seen that CH4 conversion increases and CO selectivity decreases monotonically, as the reaction pressure is varied from 3 to 9bar, indicating that the membrane permeation dominates the reaction performance. To avoid carbon deposition, steam-to-carbon molar ratio higher than 2.5 is necessary [21], high steam addition is also favor of the methane reforming and CO shift reactions. The presence of large amount steam, however, may dilute H2 in the reaction zone and thus hinder H2 permeation. In the range of 2.5 to 3.5, CH 4 conversion rate grows slowly and CO formation can be suppressed efficiently, as shown in Fig. 4(c).

3.3. Gas turbine cooling model High-temperature gas turbine performance levels are very sensitive to blade cooling requirements. In the present study, a closed-loop steam cooling (CLSC) solution is selected. The needed coolant is extracted from the high pressure steam turbine outlet, the remaining turbine exhaust steam is returned to the HRSG for reheating. After cooling the stationary and rotary hot components, the steam reaches, in practice, the reheating temperature. It is then mixed with the reheated steam from the HRSG and introduced into the intermediate pressure steam turbine section for expansion. To analyze the global performance of the cycle under investigation, a discrete (rather than differential field) model is used because of its computational convenience. The cooled turbine model with CLSC presented in previous study [22], and the refined versions of the model by Louis et al. [23] and Horlock et al. [24], is considered. It considers the turbine stage by stage, and estimates the cooling flows necessary for the stator and rotor at each stage. The stator flow is assumed to exchange heat with the main gas flow prior to flowing through the turbine, i.e., the heat exchange happens before power extraction. The rotor coolant flow cools the main stream at the rotor exit (after power extraction). For each cooling step, the required coolant mass flow is calculated as:

mc mg

0.0156

(Tg Tb ) C pg (Tb Tc ) C pc

,

(1)

where subscripts g and c refer to the main gas stream and the coolant stream, respectively. refers to blade cooling efficiency. For an advanced power generation gas turbine system, commonly used value for is 0.3. The symbol Tb refers to the turbine blade metal temperature; its value in this study is 1123K (850°C) and is kept constant in the calculation, which is validated by calibrating the model against the published performance data in [22]. The turbine in the cycle mentioned in this paper is divided into 4 stages assuming equal enthalpy drops and the first 2 stages are cooled.

140

4. System performance analysis and comparison 4.1. Performance criteria The thermal efficiency of the system is defined as: Wnet Wnetl , (2) th Q f Qsol m f LHV Qsol where Wnet is the system net power output, LHV is the fuel low heating value input, Qsol is the absorbed solar heat. Since the system input resources involve the methane chemical exergy and solar thermal energy, which is different in their energy qualities, exergy efficiency is more suitable than energy efficiency for the system performance evaluation. Assuming that methane chemical exergy is approximately equal to 1.04 times its lower heating value LHV, and the solar thermal exergy corresponds to the maximal work availability between solar collector temperature Tsol and ambient temperature T0 , i.e., Qsol·(1-T0/Tsol). Therefore, the definition of system exergy efficiency is given as follows: Wnet Wnet . (3) e E f Qsol (1 T0 Tsol ) 1.04m f LHV Qsol (1 T0 Tsol ) The contribution of the low/mid temperature level solar heat can be measured by its share in the system total energy input: Qsol Qsol X sol . (4) Q f Qsol m f LHV Qsol To evaluate the performance of the solar heat conversion in the proposed system, the net solar-toelectricity efficiency based on reference [3] is defined as: Wnet Wref Wnet Q f th , ref , (5) sol Qrad Qrad where Wref = Qf · th,ref is the net power output generated by a reference system with the same natural gas input. Here, a conventional natural gas fired gas-steam combined cycle power plant (CC) and a CC system with CO2 separation from the exhaust gas (CC-Post) are chosen as reference systems. The fossil fuel saving levels in comparison with the reference power plant, for generating the same amount of electricity, is defined as the fossil fuel saving ratio: Wnet th , ref Q f Q f th , ref . (6) SR f 1 Wnet th , ref Wnet

4.2. System performance and discussions 4.2.1 Overall performance comparison and discussion Using the computational assumptions and models given in section 3.1, LEHSOLCC, CC and CCPost systems are simulated on the same basis. The main process stream data for LEHSOLCC are shown in Table 2 and the thermodynamic performance of the three systems are summarized and compared in Table 3. It is observed that the solar heat introduced in LEHSOLCC cycle contributes 28.2% of the system total energy input, leading to a significantly increase in net power output. 31.2% and 16.4% reduction of fossil fuel input is obtained for producing the same amount of electricity in comparison with the CC-Post and CC systems, respectively. The specific CO2 emission is 25 g/kWh, lower by 141

38% than that emitted by the CC-Post system at the same level of CO2 capture rate (90%) due to the more net power output. Compared with a CC system at the same technical level and without CO2 capture, the hybrid system has a lower thermal efficiency by 8.4%-points. The thermal efficiency penalty is due to not only power consumption for CO2 capture and compression, but also the solar heat input, since the solar-alone system at the same collecting temperature level has a much less thermal efficiency. The second law efficiency is obviously more proper to evaluate energy system with multiple energy resources. Solar contribution at lower temperature is much lower from the view point of the second law efficiency. Evaluated by the exergy efficiency, the hybrid system is even comparable with the CC system without CO2 capture, embodying advantages of system integration and cascade utilization of multiple energy resources. Compared with a CC-Post system, the hybrid system exhibits better performance in terms of both thermal efficiency (by 2%-points) and exergy efficiency (by10%-points). Table 2. Main stream states of the LEHSOLCC system (points refer to Fig.3) point

t (°C)

p (bar)

m (kg/s)

1 2 3 4 5 7 9 11 12 13 16 17 20 21 24 26 27 28 29 32

15 421.8 15 74.2 15 15 180.1 530 550 550 142.8 142.8 399.2 35 122.4 1300 614.4 100.1 36.2 560

1 16.6 5 10.06 2 10.59 10.06 9.76 9.27 4.6 9.09 4.51 20.75 9 20.75 16.268 1.05 1.01 0.06 111

1.097 1.097 0.02 0.02 0.045 0.08 0.08 0.1 0.091 0.01 0.091 0.01 0.01 0.055 0.004 1.111 1.111 1.111 0.191 0.191

Percent molar composition (%) N2

O2

77.3 77.3

20.74 20.74

CH4

H2 O

CO2

CO

1.01 1.01

0.3 0.03

22.2 0.4

100 100 100 77.8 56.7

35.5

0.9

0.4

56.7

35.5

0.9

0.8 3.8

0.1 0.1 13.31 13.31 13.31 100 100

81.8 18.4 0.3 0.3 0.3

1.9 8.9

H2

Ar 0.92 0.92

100 100

72.37 72.37 72.37

13.16 13.16 13.16

6.6 100 6.6 100 100 15.3 68.8 0.86 0.86 0.86

Table 3. Systems performance comparison Items

LEHSOLCC

CC-Post

CC

Methane exergy input [kJ/mol - CH4] Low temperature solar heat input [kJ/mol - CH4] Low temperature solar exergy input [kJ/mol - CH4] Middle temperature solar heat input [kJ/mol - CH4] Middle temperature solar exergy input [kJ/mol - CH 4] Steam/methane molar ratio Solar thermal share [%] Solar to power efficiency [%] reference to CC-Post Solar to power efficiency [%] reference to CC Fossil fuel saving ratio [%] reference to CC-Post Fossil fuel saving ratio [%] reference to CC CO2 removal rate [%] Specific CO2 emission [g/kWh]

830.2 126.2 43.4 189.7 121 3.5 28.2 36.4 19.2 31.2 16.4 91 25

830.2 90 40.3

830.2 331.3

142

CO2 compressor [kJ/mol-CH4] Net power output [kJ/mol-CH4] Exergy efficiency [%] Thermal efficiency [%]

13.5 576.7 58 51.6

13 396.5 47.8 49.4

481.7 58 60

4.2.2 Parametric analysis A sensitivity analysis of some key reaction parameters has been performed to quantitatively illustrate their effects on system performance. Figures 5 to 7 depict the effect of reaction temperature (Fig. 5), reaction pressure (Fig. 6) and steam-to-carbon molar ratio (Fig. 7) on solar thermal share, solar-to-electricity efficiency, fossil fuel saving ratio, system exergy efficiency and CO2 capture rate, respectively. As mentioned before, in the 450~550oC temperature range, an increase in reaction temperature monotonically elevates CH4 conversion, leading to an increasing demand of solar heat input and saving of fossil fuel. CO2 capture rate increases as well because of higher CH 4 conversion. The energy consumption for CO2, H2 and syngas compression also increases. On the whole, the influence of reaction temperature on system efficiencies appears to level off. Solar thermal share Xsol(%) Fossil fuel saving ratio SRf(%) 100

Solar-to-electric efficiency Exergy efficiency

80

e

(%)

sol

(%)

CO2capture rate(%)

60

40

20

0 450

475

500

525

550

Reaction temperature ( C)

Fig. 5. Influence of reaction temperature on system performance The effect of reaction pressure on system performance exhibits the similar trends as that of reaction temperature as shown in Fig. 6(a). As mentioned before, higher pressure on one side hinders the reforming reaction, it on the other side boosts membrane permeation, and the later dominates in the present study, an enhancement of CH 4 conversion rate is thus observed. In case of enhanced steam pressure, as shown in Fig. 6(b), the low-temperature solar heat input diminishes due to a decrease in vaporization latent heat; the middle-temperature one for the reforming reaction predominately increases, leading to the increase of overall solar input share. Solar thermal share Xsol (%)

Middle temperature solar heat input (kJ/mol-CH4)

Fossil fuel saving ratio SRf (%) Solar-to-electric efficiency Exergy efficiency

80

e

sol

Low temperature solar heat input (kJ/mol-CH4) (%)

(%)

9

Reaction pressure (bar)

100

CO2capture rate(%)

60

40

20

7

5

3 0 3

5

7

Reaction pressure (bar)

9

0

143

50

100

150

200

250

Solar heat input (kJ/mol-CH4)

300

(a) (b) Fig. 6. Influence of reaction pressure on system performance At temperature of 550oC and pressure of 9bar, an enhancement in CH4 conversion and CO2 capture rate are observed as more reforming water is brought into the system. Solar thermal energy input increases in both water evaporation section and reforming section. Most of the imported water will be discharged in the subsequent condensation process, the working fluid mass flow rate increases slightly and yields an augment of power output. Solar-to-electricity efficiency and exergy efficiency drop slightly owning to a large solar heat input. It is noteworthy that higher steam-to-carbon ratio is not applicable, because the presence of large amount of steam causes a decline in H2 partial pressure in the reaction zone, thus weakens reaction and system performance. However, a steam-tocarbon molar ratio more than 2 is required to avoid carbon deposition.

100

580

60

CO2

560

e

sol

540

Xsol , SRf

40

20

sol

SRf Xsol

Wnet (kJ/mol-CH4)

80

e

, CO2

(%)

Wnet

520

0

500 2.5

3.0

3.5

Steam-to-carbon molar ratio

Fig. 7. Influence of steam-to-carbon molar ratio on system performance

5. Technical considerations The hybrid power system proposed in this paper has two input resources: fossil fuel and solar heat. Solar radiation, however, is not a continuously available resource like a fossil fuel. In most of the cases, a thermal storage system or a fossil fuel backup system is needed to prolong the operation hour. In the proposed system, rather than fuelling the power block directly, the solar heat is used at low/mid temperatures to contribute to the production of syngas that can then be burned to produce heat at the high temperature for high efficiency power generation. Advantages of this concept are that it has storage capability for solar energy chemically (rather than thermally) and physical independence of the solar block to the power system, i.e., the solar assistant chemical conversion can be processed separately in the most suitable site. To ensure a stable operation, more than one solar part can be adopted to match one power block. In this case, a syngas storage subsystem is essential instead of a conventional solar heat storage system. In addition, the combined cycle can run with regular natural gas when solar heat is not available. A parabolic trough direct steam generation collector (DSG) may be used to provide heat at ~200 oC for water evaporation in order to eliminate the costly synthetic oil, intermediate heat transport piping loop and oil-to-steam heat exchanger. For middle temperature solar heat collection at 500 oC, solar tower technology is considered for the thermo-chemical solar-fuel conversion. Because of the instability and discontinuousness of solar radiation, a dynamic analysis is definitely of major importance, which evaluates performance of hybrid power plants and the compares with fossil fuelled ones from a dynamic point of view, considering the daily and seasonal variability of the 144

solar source and referring the results to a yearly basis. This paper focused primarily the system integration concept and the design point performance.

6. Conclusion A novel solar-assisted hybrid system integrated with methane steam reforming and membrane separation (LEHSOLCC) has been proposed and investigated. The system mainly consists of a gassteam combined cycle, a solar-driven membrane reformer and CO 2 absorption components. The introduction of membrane reformer breaks the limitation of temperature on reaction conversion rate and achieves a nearly full methane conversion at middle temperature. Integrated with energy conversion, CO2 convergence and capture is accomplished with low energy penalty. By providing heat to the endothermic methane steam reforming, solar heat collected at 550oC is converted into reformed syngas chemical energy; and then released as high-temperature thermal energy in the advanced combined cycle system for power generation, thus achieving a highefficiency heat-power conversion. The produced pure H2, collected at the permeate side of the membrane reactor, fuels a gas/steam combined cycle. The leftover CO 2 enriched gas in the retentate zone is processed with pre-combustion decarbonization. To reduce the exergy destruction, special attention is paid to the thermal match of the internal heat recuperation, as well as to the thermochemical match between the solar heat and the reforming process. The system is simulated and compared with a conventional natural-gas fired gas-steam combined cycle power plant (CC), a CC with CO2 capture from exhaust (CC-Post). The results show that with 91% CO2 captured, the specific CO2 emission in the hybrid system is 25 g/kWh, lower than that in the CC-Post cycle by 36%. Exergy efficiency of 58% and thermal efficiency of 51.6% can be obtained, 10.2%- and 2.2%-points higher than the CC-Post cycle, respectively. Fossil fuel saving ratio of 31.2% is achievable with a solar thermal share of 28.2%, and the net solar-to-electricity efficiency, based on the gross solar heat incident on the collector, is about 36.4% compared with the CC-Post cycle with equivalent CO 2 removal rate. The effects of some key parameters on system performance have also been investigated.

Acknowledgments The authors gratefully acknowledge support of the Chinese Natural Science Foundation Project (No. 51076152) and the National Key Fundamental Research Project (No. 2010CB227301).

Nomenclature DNI direct solar radiation, W/m2 E exergy, kW LHV methane low heating value input, kJ/kg mf methane mass flow rate, kg/s Q heat, kW SRf fossil fuel saving ratio T temperature, oC Wnet net power output, kW Xsol solar thermal share Greek symbols solar collector efficiency col system exergy efficiency e 145

sol th

net solar-to-electricity efficiency system thermal efficiency

Subscripts and superscripts a Air f Fossil fuel ref Reference system rad Radiation sol Absorbed solar heat 0 Ambient state

References [1] Lewis NS., Toward cost-effective solar energy use. Science 2007; (315): 798-801. [2] Jin H., Lin R., Cascade energy utilization and gas turbine integrated energy system. Beijing: Science Press; 2008. [3] Hong H., Jin H., Ji J., Wang Z., Cai R., Solar thermal power cycle with integration of methanol decomposition and middle-temperature solar thermal energy. Solar Energy 2005; 78: 49-58. [4] Tamme R., Buck R., Epstein M., Solar upgrading of fuels for generation of electricity. ASME Trans. Journal of Solar Energy Engineering 2001; 123: 160-3. [5] Zhang N., Lior N., Use of low/mid-temperature solar heat for thermochemical upgrading of energy, part I: application to a novel chemically-recuperated gas-turbine power generation (SOLRGT) system. ASME J. of Engineering for Gas Turbines and Power (in press), 2012. [6] Li Y., Zhang N., Cai R., Parametric sensitivity analysis of an SOLRGT system with the indirect upgrading of low/mid- temperature solar heat. Applied Energy (in Press), 2012. [7] Zhang N., Lior N., Luo C., Use of low/mid-temperature solar heat for thermochemical upgrading of energy, part II: a novel zero-emissions design (ZE-SOLRGT) of the solar chemically-recuperated gas-turbine power generation system (SOLRGT) guided by its exergy analysis. ASME J. of Engineering for Gas Turbines and Power (in press), 2012. [8] Luo C., Zhang N., Zero CO 2 Emission SOLRGT Power System. ECOS2011: Proceedings of the 24th International Conference on Efficiency, Cost, Optimization, Simulation, and Environmental Impact of Energy Systems; 2011 July 4-7; Novi Sad, Serbia. Energy (in press), 2012. [9] Chen Y., Wang Y., Xu H., Xiong G., Efficient production of hydrogen from natural gas steam reforming in palladium membrane reactor. Applied Catalysis B: Environmental 2008; 80: 28394. [10] Chen Y., Wang Y., Xu H., Xiong G., Integrated one-step PEMFC-grade hydrogen production from liquid hydrocarbons using Pd membrane reactor. Ind. Eng. Chem. Res. 2007; 46(17): 5510-5. [11] Jordal IK., Bredesen R., Kvamsdal HM., Bolland O., Integration of H2 -separating membrane technology in gas turbine processes for CO 2 capture. Energy 2004; 29: 1269-78. [12] Iulianelli A., Manzolini G., De Falco M., Campanari S., Longo T., Liguori S., Basile A., H 2 production by low pressure methane steam reforming in a Pd-Ag membrane reactor over a Nibased catalyst: Experimental and modeling. International Journal of Hydrogen Energy 2010; 35: 11514-24. [13] Aspen Plus®. Aspen Technology, Inc., Version 11.1[EB/OL]-Available at:< http://www.aspentech.com/> [accessed 5.1.2009]. 146

[14] Myers DB., Ariff GD., James BD., Lettow JS., Thomas CE., Kuhn RC,. Cost and performance comparison of stationary hydrogen fuelling appliances. Arlington, VA: National Renewable Energy Laboratory, DTI; 2002 Report No.: NREL/CP-610-32405. [15] Zarza E., Rojas M., González L., INDITEP: The first pre-commercial DSG solar power plant. Solar Energy 2006; 80: 1270-6. [16] Solúcar et al., 10MW solar thermal power plant for southern Spain. 2006 Nov. Technical Project No.: NNE5-1999-356. [17] Lozza G., Chiesa P., Natural gas decarbonization to reduce CO2 emission from combined cycles. Part 1: partial oxidation. ASME Journal of Engineering for Gas Turbines and Power 2002; 124: 82-8. [18] Krzysztof L., Andrzej Z., Comparative analysis of energy requirements of CO2 removal from metallurgical fuel gases. Energy 2007; 32: 521-7. [19] Alie C., Backham L., Croiset E., et al., Simulation of CO2 capture using MEA scrubbing: a flow-sheet decomposition method. Energy Conversion and Management 2005; 46: 475-87. [20] Gas turbine world 2010 handbook. USA: Pequot Publishing, Inc.; 2010. [21] Ertesvåg IS., Kvamsdal HM., Bolland O., Exergy analysis of a gas turbine combined-cycle power plant with pre-combustion CO 2 capture. Energy 2005; 30(1): 5-39. [22] Jørgensen SL., Nielson PEH., Lehrmann P., Steam reforming of methane in a membrane reactor. Catal. Today 1995; 25:303-7. [23] Sanjay., Singh O., Prasad BN., Influence of different means of turbine blade cooling on the thermodynamic performance of combined cycle. Applied Thermal Engineering 2008; 28: 231526. [24] Louis JF., Hiraoka K., El-Masri MA., A comparative study of influence of different means of turbine cooling on gas turbine performance. GT1983: ASME international gas turbine conference; 1983 Mar 27; Phoenix, AZ, USA. ASME Paper: 83-GT-180. [25] Horlock JH., Watson DT., Jones TV., Limitation on gas turbines performance imposed by large turbine cooling flows. ASME J. of Engineering for Gas Turbines and Power 2001; 123: 487-94.

147

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

Low Exergy Solutions as a contribution to climate adapted and resilient power supply Stefan Gößling-Reisemanna, Thomas Blötheb a

b

University of Bremen, Bremen, Germany, [email protected]

University of Bremen, Bremen, Germany, [email protected]

Abstract: The focus of our research within the climate adaptation project nordwest2050 is on how the energy system in the metropolitan region Bremen-Oldenburg can be developed in order to maintain system services under climate change conditions and further uncertain and turbulent boundary conditions. To answer this question, we examined and evaluated currently emerging and already existing technologies. We conducted an innovation potential analysis and identified two promising innovative fields, which we call „Low Exergy Solutions“, and „Resilient Energy Infrastructures“. Low Exergy Solutions are characterized by the use of low exergetic ambient energy sources, or residual- and waste streams and their contribution to various energy services. Low Exergy Solutions are thus systems of matching technologies consisting of three components: an exergy source, a conversion technology and an energy service being delivered. The technologies are using waste heat and sources of material residues not currently used, thereby linking consumers and producers in regional and local networks contributing to a decentralized energy system. As a trivial example, waste heat sources with low temperature can be used for nearby space heating and other services like drying. Heat sources with higher temperatures can be used for electricity production or cooling, also across medium distances. Geothermal reservoirs can be used for local cooling and air conditioning. Results from our innovation potential analysis further show that Low Exergy Solutions contribute to climate adaptation in several ways, e.g. by decreasing the load on the electricity grid, or by providing alternatives for vulnerable energy imports. They could also make use of climate change opportunities, like increased solar heat potentials. By enlarging the diversity of the overall energy supply and by adding flexibility and redundancies, they also increase the resilience of the energy supply system.

Keywords: Climate adaptation, Resilience, Exergy, Renewable Energy

1 Introduction The results described here are based on work done within the climate adaptation project nordwest2050 in the Northwest of Germany1. The original task was to find innovations in the energy supply system that address the climate change impacts on the system. Very soon in the process it became obvious that climate change is only one of the major uncertainties that the German energy system, and with it the Northwestern subsystem, will be experiencing in the future. The German government’s decision to abandon nuclear power and its goal to have 35% of gross electricity consumption and 18% of gross energy consumption from renewable production by 2020 [1] currently alters the German energy supply system considerably. The common denominator of the coming climate change impacts and the current turbulent dynamics of the energy transition are the uncertainties surrounding them. It thus seemed wise, to not just search for innovations in the energy supply system that address climate change, but to develop design guidelines for an energy supply system that is better equipped to cope with turbulence and uncertainties. The research setup of our project was then adjusted and now incorporates 1

See http://www.nordwest2050.de

148

 a two-fold vulnerability assessment: one with the explicit focus on climate change impacts and one with a focus on the general vulnerability of the Northwest German energy supply system  an innovation potential analysis with a focus on regional innovation demands and regional potentials  the development of design guidelines for resilient energy systems, i.e. systems that maintain their system services even under strong external disturbances and internal failures  the implementation of region specific demonstrators to show the feasibility of short-term adaptation and resilience building  the participatory design of a “Roadmap of Change” for the region, not only limited to the energy sector, but also including other sectors of the economy The overall objective of the energy related activities in this project is to find design guidelines for the whole regional energy system. Naturally, individual analysis of technologies will have to be concise and geared towards the integration into the existing energy system and towards structural changes in the current setup. The project nordwest2050 therefore encompasses analysis and design on multiple levels of the energy system: supply chains, regional governance systems, societal needs and perceived risks, environmental impacts, conflicts with other sectors, and, of course, technologies. The results presented here should be seen in this context: as part of a larger regional transformation process which includes technological aspects as much as innovation aspects and aspects of economic feasibility and social acceptance. In this paper, we will mainly present results from the vulnerability assessment, its consequences for the adaptation demands and innovation needs in the region, the concept of resilient energy systems and present one of the identified innovation fields for climate adaptation and resilient energy systems which we call “Low Exergy Solutions”. The aim of this paper is to show how climate adaptation, building resilient energy systems and climate mitigation can simultaneously be addressed by a set of technologies chosen on the basis of scientific analysis and a guiding concept inspired by natural systems.

2 Resilient energy systems Our understanding of a resilient energy system is based on the work by ecosystem theorists Holling, Gundersson and others [2][3][4]. There is no unique definition of resilience, so we adopted an ecosystem based definition from [5] Resilience „reflects the capacity (i. e. the underlying mechanisms) of [eco]systems to maintain service in the face of a fluctuating environment and human perturbation” [3][4][5] For socio-technical systems we translated this into our definition of resilience: Resilience describes the ability of a system to maintain its services under stress and in a turbulent environment, i.e. even in the face of massive external perturbations and internal failures [6]. In the theoretical discussion of the resilience concept, we identified several design elements of socio-technical systems that help increase their resilience [7]: 149

Table 1. Design Elements for resilient socio-technical systems (abstract level) System capabilities System resources availability System structure  Adaptability  Resistance  Creativity and Design

 Energy and material  Information  Finances

 Diversity  Redundancy and modularity  Balance of positive and negative feedback mechanisms  Buffers and storage  Dampers and attenuators

These design elements can be applied to energy supply systems to direct innovations towards a more resilient design of such systems. Resilience in this sense is used as a “Leitkonzept” (guiding concept). The rather abstract design elements can further be broken down into more system specific, e.g. “Creativity and Design” translates to “Open Interfaces” in the case of resilient energy systems, meaning the ability of a resilient system to allow the integration and flexible combination of several energy carriers and infrastructure systems (gas, water, district heat, etc). The set of energy specific design elements constitutes the “Gestaltungsleitbild” (design guiding concept) of what we call resilient energy systems. The guiding concept should not be taken as an analytical metric. It rather points the search for innovation into a specific direction. To give an example, the diversity of an energy system is a function of the diversity of the resources it is able to use (resource diversity) as well as the diversity of the technologies comprising the system. On the resource side, a system is more resilient, the more resources it can access. Thus, innovations that enable the system to access currently unused or underutilized resources, increase the resilience of the system. The full evaluation of an innovation is however not limited to the resilience increasing effects. Other factors, like environmental risks, economic feasibility etc. have to be taken into account (see section 4). More on the use of guiding concept of resilience can be found in [7].

3 Vulnerability of the energy supply system in Germany’s Northwest In order to determine innovations enabling the adaptation of the energy supply system in Germany’s Northwest to climate change, a first step was to assess the vulnerability of the system. As mentioned in the introduction, it soon became obvious that climate change is only but one of the many changes that the energy supply system is facing, all characterized by high uncertainties and possible elements of surprise. In order to analytically capture the vulnerability related to these uncertainties and surprise, we altered our vulnerability assessment by distinguishing between climate related vulnerabilities and structural vulnerabilities. Climate change vulnerabilities correspond to potential impacts from climate change, while structural vulnerabilities correspond to weak spots in the architecture of the system that make it more vulnerable to internal failures or external shocks. Without going into the details2, a typical example for a structural vulnerability is the missing mechanism for dealing with conflicts resulting from an increasing share of renewable in the electricity sector. While e.g. the pressure on land-use in the biomass sector, the conflicts surrounding wind park sites and the issues around the intermittency related grid-stability problems are increasing, there is no consensus and no clear regulation in sight to deal with these problems. This is an architectural flaw which currently poses a threat to the stability and the future pathway of the German energy supply system in general, not just in the Northwestern parts. 2

See [8] and [9] for a detailed discussion on the vulnerability assessment methodology.

150

The results of the vulnerability assessment (climate change related and structural) are summarized in Table 2. To understand the results, it is to be noted that high potential impacts and high adaptive capacity cancel each other out, while medium potential impacts can still lead to high vulnerability if adaptive capacity is low. Table 2. Summary of results from the vulnerability assessment of the energy supply system in the Northwest of Germany. DSM = demand side management. Primary energy Coal

Gas

Wind

Grid-bound energy / Demand / Applications distribution Biomass Electric. Gas Heat Cooling DSM

Pot. impacts (climate)

Low

Low

Low

Medium Medium Medium Medium Low

pot. impacts (stuctural)

Medium Medium Medium High

Adaptive capacity

Medium Medium High

Medium Medium Medium Medium Medium High

Climate Vulnerability

Low

Medium Medium Medium Medium Low

Structural Vulnerability

Medium Medium Low

Low

Low

High

High

High

Medium

Medium Medium Medium Medium

Low

Medium Medium Medium Low

The structural vulnerability is generally higher than the climate change related vulnerability. One has to bear in mind though, that due to the chosen definition of structural vulnerabilities, they in some cases include vulnerabilities which result from climate change. From further analysis of the energy supply and electricity generation in the region, we deduced another structural shortcoming of the region’s energy supply system: it only has a comparatively low diversity [25]. As a conclusion from this assessment, we identified several priority fields for innovation in the regional energy system:  Reduce sensitivity, especially by providing long-term political and legal framework (biomass and electricity)  Install conflict management tools (biomass)  Increase resource diversity (electricity, gas, heat, cooling)  Increase share of renewables (electricity)  Increase storage capacity, strengthen networks and load management (electricity)  Couple heat sources and cooling services (cooling, DSM) These recommendations for innovation were the input for a subsequent innovation potential analysis, where regional potentials and demands where combined to derive concrete innovation candidates.

151

4 Innovation potential analysis The methodology for the innovation potential analysis (IPA) has been developed within the project nordwest2050 (see [10]). It consists of four stages: 1. System definition: regional scope and cluster definition (here: the energy sector) 2. Identify innovation fields from vulnerability assessment, climate adaptation debate, innovation trends, and guiding concepts (here: resilient energy supply systems) 3. Describe the innovation system, identify and assess innovation capabilities of the regional sector 4. Identify innovation candidates and evaluate based on criteria list, choose candidates for demonstration projects The regional scope is given by the metropolitan region Bremen-Oldenburg, a European metropolitan region consisting of counties and cities in the Northwest of Germany around the two larger cities of Bremen and Oldenburg (see www.frischkoepfe.de). Within the region, the innovation potential is mainly determined by the regional energy sector and the resources available. On the conventional side, the energy sector is characterized by two large utility companies (EWE AG and swb AG) and further utility companies operating power plants in the region. The electricity generation is currently dominated by hard coal powered conventional power plants, with wind, biomass and solar gaining quickly. The renewable generation, currently mainly wind and biomass, contributes approx. 29% to the electricity being produced in the region with 20% alone from wind and 7% from biomass (biogas plants).

4.1 Innovation needs There are a few climate change driven innovation necessities for conventional power plants, mostly regarding cooling water availability. However, with the current regulation regime, there are hardly any new power plants being built without additional cooling towers, so that the regional cooling water availability will not pose a serious problem in the future3. The wind generation is hardly affected by climate change, except for the increasing likelihood of severe storms, but the climate data on this is inconclusive. The main challenges for the regional energy supply system come from structural problems, as was derived in the vulnerability assessment: especially biomass and electricity generation and distribution are affected by the uncertain regulatory and political framework and by the conflicts around the increasing renewable production and the extension of the grid. The heat market (mainly natural gas and district heating) is affected by the decreasing number of heating degree days, which makes the extension of grids and the addition of new capacity economically less attractive. For cooling, the situation is somewhat reversed, since regional average summer temperatures are expected to rise by 1-2 degrees Celsius until 2050 and heat waves increase significantly in number, duration and temperature level [9]. With buildings becoming ever better insulated, there is a clear demand for making better use of the increasingly unused heat being generated in the regional energy system. This especially applies to the many hundred biogas plants in the region, which are almost entirely situated in rural areas with only limited connectivity to district or near-range heating networks. One long-term solution for these plants is to separate biogas generation and electricity production by either building raw gas pipelines and satellite plants or by upgrading the raw biogas to natural gas quality and feed it into the dense natural gas grid in the region. For a short term solution, it makes sense to use the heat from biogas plants for other energy services, like cooling, drying, air-conditioning, etc.

3

This differs from other regions in Germany and Europe, see [11].

152

4.2 Innovation fields With the results from the vulnerability assessment and considering the structure of the regional energy system, with its current focus on coal, wind and biomass and a rising share in renewables to come, we have derived two promising innovation fields to increase the resilience of the system:  Low Exergy Solutions  Resilient Energy Infrastructure Both innovation fields address the issues raised by the vulnerability assessment, whether regarding climate change impacts or structural weaknesses and incorporate design elements from the theoretical discussion of resilient energy systems. In essence, they both should reduce the demand on the electricity network, especially in the summer time, provide a richer and more divers resource base, and provide flexibility and adaptability. While Low Exergy Solutions build on low exergetic and currently unused heat and material flows, resilient energy structures are based on interoperability of renewables with the grid, buffering and storage capacities and intelligent management of load and demand. In the following we will focus on the field of low exergy solutions only.

4.3 Low-Exergy-Solutions The technologies for utilizing low-exergy flows are known for most applications, but they are in different stages of development. Innovation in the use of low-exergy sources is therefore needed mainly in the appropriate connection of previously unused energy sources (such as industrial waste heat, waste water heat, geothermal cooling/heating, material residues, etc. ) with well-known conversion technologies (e.g. adsorption chillers, heat pumps, thermoelectric power generation, fermenters, etc. ) for the provision of appropriate energy services (such as industrial cooling, airconditioning, drying, power supply, biogas, etc. ). In some cases, the conversion technologies are not yet developed or in the early phase of development, like technologies for using low exergy material flows (agricultural residues, wastes, cellulose-rich substances other than wood etc.) We have scanned the literature on utilization technologies for low exergy flows and have identified a set of twelve technology components and six systemic combinations with promising potential for increasing the resilience of the regional energy system: Table 3: Possible technology components and systems within the Low Exergy Solutions innovation field. Components (sources and conversion Systems (providing services) technologies) Comb. heat and power District heating grid Biogas plant + CHP– Mobile heat Solar thermal Geothermal cooling Biogas plant – raw gas network – satellite installations CHP – absorption chillers Industrial waste heat Long term thermal Agricultural residues – biogas plant – CHP – storage absorption chillers Heat pumps Organic Rankine Cycle Bio-energy villages (raw gas grid, CHP) Ad-/Absorption Mobile heat storage Food residues – biogas plant – CHP – chillers (PCM) absorption chillers Cellulose fermenters (RuminoTech) RuminoTech – raw gas grid – CHP

153

All technologies and systemic combinations are currently assessed against a set of indicators developed in the context of the regional adaptation project. The innovation indicators include       

Status within the technology diffusion curve Climate mitigation potential Effect on vulnerability of the regional energy system Effect on the resilience of the regional energy system Plausibility of the expected effects Technical and economic risks Transferability

More indicators are currently added and final results will be published on the project website shortly (www.nordwest250.de). The pilot projects discussed below are already evaluated using the above indicators (see appendix for a graphical representation of the results).

4.3.1 Pilot projects Basically, applications of Low-Exergy-Solutions can be established in the fields of private households, commerce and industry, and also in the public sector. We tried to derive concepts for flagship projects with broad coverage of these areas and high visibility. Since we are striving to derive options for a region wide strategy for resilient energy supply systems, it is important to engage with regional stakeholders in an early phase. The respective stakeholders then also have the best access to the regional application areas and can inform other end-users about such installations. We have started to implement three pilot projects, described below, and will continue to bring together technology developers with potential users and investors to explore the feasibility of a region wide diffusion of innovative Low Exergy Solutions. The pilot projects are accompanied by scientific analysis, e.g. measuring performance and efficiency and economic feasibility, to derive conclusions about the potential contribution of Low Exergy Solutions to a resilient regional energy supply system.

4.3.2 Pilot project 1: Cooling turkey barns and closing energy and material cycles On an agricultural farmstead in the administrative district of Vechta, the resident farmer operates a turkey breeding facility. On this farm there are 3 turkey barns with each barn being designed for 7000 turkeys. The farmer also operates a biogas plant with a 500 kWel CHP unit. The fermentation substrate of the biogas plant consists of manure from the turkeys and corn cultivated on the farm land. The biogas produced is converted into electricity by the CHP plant. The electricity is currently fed into the public grid and remunerated according to the German feed-in tariffs. The waste heat from the cogeneration units currently heats the fermentation vessel of the biogas plant, the turkey barns and the residential buildings on the farm. 4.3.2.1 Climate change considerations In the summer, with almost no heating demand in the buildings, the heat from the cogeneration unit is currently expelled to the surroundings. On the other hand, the animals in the barn demand continuous ventilation to dissipate their accumulating body heat. With the climate change noticeably changing summer temperatures, this causes serious problems for the turkey farmer, already now. According to [12] „… a room temperature of 21 °C to 23 °C is sufficient with an 154

auxiliary heating system. It's also important that the animals are not to be kept warm, since too high house temperatures can speed up the heartbeat and the breathing. This is combined with a high burden on the young tissue, so that later health problems such as abdominal dropsy or bleeding can be the consequence.“ In the summer, the rising ambient temperature can lead to an unacceptable increase of the temperature in the turkey barns and the animals are subject to heat stress. Even further rising temperatures can lead to the death of the turkeys, as has already happened in the recent past. With climate change it is to be expected that the mean air temperatures will continue to rise and an increase of heat waves must be expected. This has been confirmed by evaluating regional climate models [9]. A short term adaptation strategy, and implemented in this pilot project, is to aircondition the turkey barns. In the longer term, turkey breeding should be based on better suited living conditions, e.g. in less densely populated barns, to avoid the heat stress problem altogether. 4.3.2.2 Energy considerations From an energy perspective, this pilot project seeks to demonstrate the possibility of utilizing so far unused heat sources, thereby realizing primary energy savings and closing heat and material cycles. The heat produced by the CHP drives an absorption refrigeration system for the air conditioning of the turkey barns. Approx. 400 kW thermal capacity is available for the refrigeration plant. The heat is delivered to the absorption chiller at a temperature of 95 °C (with 50 °C in the return flow). The chiller delivers cold water at a temperature of 6° C, with 12° C in the return flow. The currently planned cooling unit will use 100 kW of the thermal energy from the CHP plant to produce 70 kW of cooling energy (COP 0.7). The actual exergy efficiency can only be calculated when the plant is operational, but based on the design parameters, and the temperature levels of the heat source and the return flow of the cooling circuit, an estimate can be derived. The exergetic service of the device is the removal of 70 kW of thermal energy from the return flow at 12° C, while the exergetic input is the heat flow of 100 kW into the system at 95° plus the electrical power consumption of about 2 kW. The rational exergetic efficiency of the absorption chiller is then around 0.15. This number could be significantly improved, if the offheat from the CHP could be delivered to the absorption chiller at a higher temperature. This, however, would incur greater investments and is currently not pursued any further. Even without further improvements, the system allows using currently unused heat sources at an affordable cost. To reduce the peak power of the absorption refrigeration system, a thermal buffer will be included. The air distribution in the stable must be carefully planned so that no "cold-" or "heat islands" be produced, which would have the risk of the animals being excessively cooled or overheated. The cooling capacity for the air conditioning system is laid out to be sufficient even till the end of the fattening period. 4.3.2.3 Regional considerations and potentials Most absorption cooling units are currently being deployed in the North America and in Asia. However, according to [23] there is a great market potential also in Germany. The food industry in Germany alone operates around 16.500 cooling units, mainly of the compression type, from which a large portion could be substituted with absorption chillers. With regard to the metropolitan region of Bremen-Oldenburg, there is some relevant potential for combining both of the low exergy sources in this pilot project on a larger scale: using manure for producing biogas and using offheat from the biogas CHP plants for heating and cooling. If all manure in the region was converted in conventional biogas plants, approximately 1700 GWh of electricity and 2600 GWh of thermal energy could be generated annually [26]. The absorption chiller in this example has a COP of about 0.7. So if only 25% of the potential thermal energy from manure driven biogas CHP plants could be used for cooling, using readily available absorption chillers as in this pilot project, cooling services in the order of 455 GWh could be provided. The German cooling demand is around 20,000 GWh per year [29]. Based on the number of people 155

living in the metropolitan region (2.7 Mio), the regional cooling demand is in the order of 660 GWh, so that approximately two thirds could be generated from this type of offheat. Compared with regular compression chillers (COP around 4), approximately 114 GWh of electrical energy could thus be substituted with offheat.

4.3.3 Pilot project 2: Geothermal cooling of a data center in Bremen The data center operator Consultix GmbH intends to build a new data center in Bremen. The data center’s waste heat must be removed from the buildings to avoid overheating of the equipment. Traditionally, this is achieved by electrically driven (compression) cooling systems. In essence, the waste heat of the servers is expelled to the ambient air by an external dry cooler. Between 50% and 150% of the electricity demand of the servers will be needed in addition in order to supply the necessary cooling [13][30]. High-quality electrical energy (100% exergy) is used to be discarded as waste heat. This is also associated with high costs, of course. 4.3.3.1 Climate change considerations With increasing cooling needs to be expected in the future due to higher average temperatures and additional heat waves, the above mentioned problem will be aggravated. When this cooling demand is covered by compression refrigeration units, an increasing burden on the electricity grid is to be expected, adding to the already increasing burden from increasing the share of renewables in the grid. Additionally, data center operators are currently competing for the “greenest” solution. Reducing the cooling demand or meeting it in a climate friendly manner is the first and most promising objective in this context. Here, Low Exergy Solutions promise considerable potential for saving energy, reducing carbon emissions, answering the increasing cooling demand from higher temperatures and heat waves and relieving the electricity grid. 4.3.3.2 Technological solution A natural choice for the cooling system seems to be free cooling. Here, cool air is directly used as a coolant. However, in the summer, at high ambient temperatures this technique does not suffice. An alternative could be cooling using groundwater. In this case problems arise from the aggressiveness of the groundwater and from the high iron content leading to the silting of the installation [13]. As a result, in this pilot project we propose to use geothermal probes. The geothermal probes are housed on the same site as the data center. Geothermal probes provide the possibility for cooling without the use of compression refrigeration units even in the summer. The system is ideally combined with free cooling: as long as the temperature is below 19 °C, the cooling demand is met by using ambient air. When the outside temperature rises beyond that, the waste heat is expelled using geothermal probes in the ground beneath the building. The geothermal probes will extend up to 150 m deep into the earth. In the summer period, approximately 3-4 months, when the ambient temperature is above 19 °C, the geothermal probes can also be used as a heat storage unit. In the winter months it is possible to cool down the soil with the geothermal probes again by using the stored heat for nearby residential buildings, for example. This pilot project is currently in the preplanning phase, where geothermal surveys and test drillings are made. 4.3.3.3 Energy considerations Since the data center is still under construction and no measurements of the cooling unit can be taken, an exergetic assessment can only be performed on design data. Towards this purpose, the air temperature in the data center is assumed to be 22° C (TD), according to specifications regarding the optimal operation of servers. The outside temperature is assumed to be 19° C (T0), which is the temperature where free cooling is assumed to not suffice to cool the data center. The heat removed from the data center is delivered to the ground at a temperature of 10° C (TG). Although the ground temperature will not stay constant during the operation, for an estimate of the exergetic efficiency 156

the assumption should be reasonable. The actual electrical power consumption of the geothermal cooling unit is expected to be in the range from 10 to 20% of the amount of heat to be removed, i.e. the COP is around 5-10. According to the supplier of the geothermal cooling unit this is a rather conservative estimate, depending of course on ground temperature, thermal conductivity, efficiency of the pumps and other factors. With this data, the exergetic service of the unit can be derived as “heat flow Q removed at a temperature of 22° C”, i.e. the exergetic service is Q(1-T0/TD) while the exergetic input is derived from Q/COP, since only electrical energy is used. The rational exergetic efficiency then is between 0.05 and 0.09, depending on the COP. This seems rather low, but it should be compared to a compression chiller, which operates at a COP between 3 and 5, resulting in rational exergetic efficiencies between 0.03 and 0.05. One could argue, however, that the environment for this cooling unit is not really the ambient air, but the ground at 150 m depths. The reference temperature then becomes TG and the rational efficiency improves markedly (to values between 0.17 and 0.3) due to the higher exergetic service at the same exergetic expense. If, in a second stage, the stored heat in the ground can be used to heat nearby houses or offices, the exergetic evaluation would have to include these additional services. This analysis will be done once the data center and the cooling unit is operational and measured data is available. 4.3.3.4 Regional considerations and potentials A typical data center uses about 50% of its electricity consumption for cooling4. In 2008, the data centers in Germany used around 10 TWh of electricity [30], of which approximately 5 TWh are for cooling only. With the help of innovative cooling technologies, like the described geothermal cooling, the efficiency of these data centers can be improved significantly. With the above mentioned increase in the COP from 3 (on average) to between 5 and 10, the electricity consumption of data centers in Germany could be lowered by 2 to 3.5 TWh, or 20-35%, respectively. For the metropolitan region of Bremen-Oldenburg there is no data on the actual numbers of data centers available. However, based on the gross regional product compared to the German GDP, resulting in a share of 3%, the number of servers in data centers in the region should be in the order of 65000 with an electricity consumption of 300 GWh (based on data in [30]). If the above savings could be realized in all of these data centers in the region, the saved electricity would be in the order of 60 to 100 GWh. One must note, however, that geothermal cooling is not available everywhere, depending on geological conditions and space requirements for the geothermal probes. On the other hand, geothermal cooling is not limited to data centers, also hotels, offices, the food industry and other sectors are potential users. The full potential for this cooling technology in the region is currently being evaluated.

4.3.4 Pilot project 3: The use of cellulose-rich substrate in biogas plants The third currently pursued pilot project extends the low exergy philosophy to biogenic materials: using low exergy material flows or residues to generate useful energy services. RuminoTech is an innovative biogas technology that mimicks the functions of a cow’s stomach system (rumen). Different from conventional plants, based on the rear end of the biological digestive tract, the RuminoTech technology is based on the front end.

4

http://www.google.com/about/datacenters/inside/efficiency/

157

4.3.4.1 Consideration of sustainability and resilience The rumen in ruminants is one of the most efficient systems in the course of evolution for the utilization of cellulose. It uses the energy of cellulose - the main component of plant cell walls - as a motor for its metabolism. For an overview of current rumen based biogas technologies, see [15]. If humans could use cellulose more efficiently for energy and material needs, some of the many resource problems could be alleviated. Currently the technology for using rumen based biogas production is challenged by unstable biological environments, by a lack of animal mucus substitutes, and by mimicking the complex flow of fluids inside biological digestive systems [15]. The RuminoTech technology follows the principle of the rumen and is believed to have overcome the mentioned challenges. The technology has been tested on the lab-scale [16] and it is currently being scaled up to the 100 kW level. The rumen based biogas technology promises to utilize plants more effectively than classical biogas plants. The renewable raw material cellulose could thus be used for renewable energy generation without many drawbacks of conventional biomass utilization. Because the fruit of the plants is not necessary for the technology to work, the conflicts between the cultivation of food and energy crops could be avoided and one of the major controversies around sustainably using biomass would be answered. A further improvement for current biogas installations could result from feeding fermentation residues (digestates) into this technology to produce usable methane, while the nutrients would remain preserved to a large part. The inventor suggests that also straw, hay, leaves, old paper or organic residue materials can be used as a fermenting material. If realized on a larger scale, this could thus add substantially to the resource diversity of the energy system and increase its resilience. The resulting high-quality biogas can be processed with customary procedures directly to natural gas or be converted with CHP plants into electricity and heat [14]. The functioning of the technology has been demonstrated on the laboratory level and a first pilot plant is currently being developed. In this pilot project, a demonstration plant with a capacity of 100 kWel will be installed in parallel to an existing biogas plant. The available biogas plant is established in close proximity to a commercial garden centre. This opens the possibility for the RuminoTech biogas plant to be tested with different fermenting materials. The biogas produced is fed into an existing CHP plant. 4.3.4.2 Energy considerations and regional potentials The first installation of this technology in the region will be run on digestate from a nearby conventional biogas plant. Based on first approximations, the RuminoTech technology can produce between 50 and 100 l of methane from 1 kg (dry matter) of digestate [22]5. The conventional biogas plant (500 kWel) is run on corn and produces around 2.5 t of digestate each day, enough to run the RumminoTech plant on 100% capacity. The weight of digestate is thereby reduced by approximately 50%, and the resulting residues have a very high lignin content, which can still be used energetically, e.g. as pellets in space heating systems. For the metropolitan region the potential in using digestate is significant. In the region of Bremen-Oldenburg 2.5 million tons of corn for biogas production are harvested each year, resulting in approximately 1.875 million tons of digestates with a dry matter content of 0.625 million tons. With the above mentioned 50 l methane per kg dry matter digestate, this would result in 3125 m³ of methane which could e.g. be converted into 130 GWh of electricity, approximately 1% of the regional electricity consumption. From a sustainability perspective, however, it is not recommendable to run the RuminoTech technology as an “afterburner” of corn based biogas plants. The RuminoTech technology offers more, especially the possibility to solve the problem between crop production for food and the production of energy. We are currently analyzing the feasibility of this concept in the regional context, determine economic and energetic efficiency, assess the flexibility of the technology and 5

Data is scaled up from laboratory results and needs to be confirmed for the full scale unit experimentally.

158

explore different diffusion scenarios in the region. An estimate of the amount of straw from different grain crops is given in Table 4. Table 4: Straw potential from different grain crops in the metropolitan region of BremenOldenburg. Data based on [18] Winter Spring Winter Spring Rye Oats Triticale total amount of Wheat wheat barley barley t/a t/a t/a straw t/a t/a t/a t/a t/a 554,317 2,186 323,878 29,163 162,515 16,808 153,333 1,208,661 The RuminoTech technology is reported to produce 400 dm³ of biogas per kg of straw [22]. This number is scaled up from laboratory experiments and has to be evaluated for the full scale unit experimentally. If the order of magnitude is correct, the straw potential in the region of BremenOldenburg could be used to produce around 480 million m³ of biogas. With an average upper heating value of 6.4 kWh/m³ and an efficiency of 40% for the conversion to electricity, the straw produced in the region could potentially generate 1.24 TWh of electricity, or 7% of the regional consumption [27]. Other potential sources of substrate need to be evaluated in order to be able to assess the potential in applying the RuminoTech technology in the region. Based on the above given description and data, the potential is quite promising, but has to be accompanied by an integrated concept for renewable energies combining the strength of different technologies.

4.4 Conclusion and outlook From our analysis so far we conclude that Low Exergy Solutions can play a significant role in adapting the regional energy system to climate change and to increase its resilience. The innovation potential assessment and the preliminary analysis of the above pilot projects substantiate this conclusion. Additionally, there is a relevant energy (savings) potential in the metropolitan region Bremen-Oldenburg when these technologies would be used. Nevertheless, at this stage of the project conclusions can only be preliminary and must be confirmed by analysis and evaluation of the currently installed pilot projects. Efficiencies, energy potentials and innovation criteria look promising. However, the real potential of Low Exergy Solutions for the region is depending on a large number of factors, including economic feasibility, regulation and permitting issues, social acceptance, robustness of the technologies, etc. We have to wait for the scientific evaluation of the pilot projects before we can derive robust conclusions, but there are several indications for Low Exergy Solutions being helpful in the context of climate adaptation and resilient energy systems:  they relieve pressure on infrastructures and resources  they diversify the energy supply  when properly integrated, they add flexibility in delivering energy services  they broaden the used resource base  they increase efficiency of renewable energy conversion systems  they consist of technologically robust components  they save high exergetic resources (electricity/fossil fuels)  they decrease carbon emissions

159

Still, there are several open questions to be answered in the pilot phase:  do they make economic sense?  are there ecological risks?  can they successfully be transferred to other sites throughout the region?  do they integrate well with existing infrastructures and technologies? The research on Low Exergy Solutions is only one approach for increasing the regional energy system’s resilience and for adapting it to climate change. In the Resilient Energy Infrastructure field for example, we also assess different storage technologies for electricity and intelligent solutions balancing demand and supply. The task of the following years will be to combine results derived from these innovations fields and organize them into a strategy for transforming the current regional energy system towards more resilience. The theoretical considerations on what constitutes a resilient system will then come back into play and have to be checked against the real-life experiences from these test cases.

4.5 References [1] German Government, The Federal Government's energy concept of 2010 and the transformation of the energy system of 2011, Available at: [accessed 10.02.2012]. [2] Holling, C. S., Gunderson L. H., Resilience and Adaptive Cycles. In L. H. Gunderson, C. S. Holling, Panarchy, 2002. [3] Carpenter et al., From Metaphor to Measurement: Resilience of What to What? Ecosystems 4(8), 765–781, 2001. [4] Folke, C. et al., Resilience and Sustainable Development: Building Adaptive Capacity in a World of Transformations. Scientific Background Paper on Resilience for the process of the World Summit on Sustainable Understanding transformations in human and natural systems. Washington: Island Press. Development on behalf of the Environmental Advisory Council to the Swedish Government, 2002. [5] Brand, F., Ecological resilience and its relevance within a theory of sustainable development. UFZ Centre for Enviromental Research, Leipzig, 2005. [6] Fichter, K., Gleich, A. v., Pfriem, R., Siebenhühner, B., Theoretische Grundlagen für erfolgreiche Klimaanpassungsstrategien. Nordwest2050 Berichte Heft 1. Bremen/Oldenburg: Projektkonsortium, nordwest2050, p23, 2010 [7] Stührmann S, Gleich A von, Brand U, Gößling-Reisemann S, Mit dem Leitkonzept Resilienz auf dem Weg zu resilienteren Energieinfrastrukturen. In: Decker M, Grunwald A, Knapp M, editors. Der Systemblick auf Innovation - Technikfolgenabschätzung in der Technikgestaltung. Berlin: edition sigma, 2012. [8] Fichter, K., Gleich, A. v., Pfriem, R., Siebenhühner, B., Theoretische Grundlagen für erfolgreiche Klimaanpassungsstrategien. Nordwest2050 Berichte Heft 1. Bremen/Oldenburg: Projektkonsortium, nordwest2050, chapter 2, 2010 [9] Schuchardt, B., Wittig, S., Spiekermann, J., Klimaszenarien für ‚nordwest2050’ Teil 2: Grundlagen. Nordwest2050 Werkstattbericht Nr. 3. Bremen/Oldenburg: Projektkonsortium, nordwest2050, 2010 [10] Fichter, K., Hintemann, R., Leitfaden Innovationspotenzialanalyse. Nordwest2050 Werkstattbericht Nr. 5. Bremen/Oldenburg: Projektkonsortium, nordwest2050, 2010 160

[11] Rothstein, B.; Parey, S.: Impacts of and adaptation to climate change in the electricity sector in Germany and France. In: Ford, J. D.; Berrang-Ford, L. (Hrsg.): Climate Change Adaptation in Developed Nations - From Theory to Practice. Heidelberg, 2011. [12] Henk Ten Haaf, Putenmast: Produktionstechnische Tips. Dtsch. Geflügelwirtschaft und Schweineproduktion (DGS) 23, 35-37, 1992 [13] Dickehut, A., General Manager, Consultix GmbH, Informations-Technology, statement 07.02.2012 [14] Strecker, M., Rumen based process for the sustainable production of biogas. Presentation at GIDIRAT 2010, Hannover, Germany. 2010. Available at http://www.isah.unihannover.de/pages/aktuelles/gidirat/downloads/19_Strecker.pdf [accessed 19.05.2012] [15] Bayané. A., Guiot, S.R., Animal digestive strategies versus anaerobic digestion bioprocesses for biogas production from lignocellulosic biomass. Reviews in Environmental Sciences and Biotechnology 10:43–62, 2011 [16] Stopp, P., Weichgrebe, D., Rosenwinkel, K.-H., Strecker, M., Breves, G., DAUMENEnergy “Design for separation and augmented methanisation of fibres substrates – contribution to sustainable biogas production“, in: Proceedings of Biogas Science 2009, Erding, Germany. Schriftenreihe der Bayerischen Landesanstalt für Landwirtschaft. 2009 [17] Gößling-Reisemann S, Gleich A von, Stührmann S, Wachsmuth J, Climate change and structural vulnerability of a metropolitan energy supply system – the case of Bremen-Oldenburg in Northwest Germany. Journal of Industrial Ecology (submitted, in review) [18] Helmich, J., Thermo-chemische Vergasung landwirtschaftlicher Biomassen Potenzialanalyse und Beitrag zu einer resilienten Energieversorgung in der Metropolregion Bremen-Oldenburg (Diploma thesis). Bremen, Germany: University of Bremen, 2010. [19] Gers-Grapperhaus, C., Biogas-Aktuelle Situation in Niedersachsen. Presentation at OLECMeeting; 1 June 2011. Available at http://www.energiecluster.de/files/christoph_gersgrapperhaus_-_biogas_aktuelle_situation_in_niedersachsen_1.pdf [accessed 28.4.2012] [20] Berliner Energieagentur GmbH, Machbarkeitsstudie thermische vs. konventionelle Kälteerzeugung am Bürogebäude Tiefstack der Vattenfall Europe Hamburg, Berlin, 2008. [21] solarnext, Data leaflet „chilli Cooling Kit WFC 70”, available at http://www.solarnext.eu/pdf/ger/products/111212_chillii_kit_WFC70_d.pdf [accessed 18.05.2012] [22] Riedel, O., CEO of Victeos Group and RuminoTech GmbH: oral communication, 09.05.2012 [23] Zürich, S., Beim Kühlen bis zu 80 Prozent Primärenergie einsparen, BHKS-Almanach, 2011 [24] Landwirtschaftliches Zentrum für Rinderhaltung, Grünlandwirtschaft, Milchwirtschaft, Wild und Fischerei Baden-Württemberg (LAZBW), Einsatz von Wirtschaftsdüngern, available at http://gruenland-online.de.dedi335.yourserver.de/html/duengung/wirtschaftsduenger/naehrstoffanfall/naehrstoffanfall.html [accessed 18.05.2012] [25] Wachsmuth, J.; Gleich, A. von; Gößling-Reisemann, S.; Lutz-Kunisch, B.; Stührmann, S.: Sektorale Vulnerabilität: Energiewirtschaft. In: Schuchardt, B.; Wittig, S. (Hrsg.): Vulnerabilität der Metropolregion Bremen-Oldenburg gegenüber dem Klimawandel (Synthesebericht). nordwest2050-Berichte Heft 2, Projektkonsortium ‚nordwest2050’. S. 95112. Bremen/Oldenburg. 2012 [26] Hammerschmidt, A. et. al., Potenzialanalyse für Strom und Wärme aus erneuerbaren Energien in der Metropolregion Bremen – Oldenburg im Nordwesten (Project report), University of Bremen, 2011 161

[27] Eickemeyer, T. , Strombilanz der Metropolregion Bremen-Oldenburg (Diploma thesis), Bremen, Germany: University of Bremen, 2011 [28] Deutscher kältetechnischer Verein (DKV), Final report „Energiebedarf bei der technischen Erzeugung von Kälte“. Hannover, Germany, 2002. [29] Bettgenhäuser, K., Boermans, T., Offermann, M., Krechting, A., Becker, D., Kahles, M., Pause, F., Müller, T., Klimaschutz durch Reduzierung des Energiebedarfs für Gebäudekühlung. Umweltbundesamt Berlin, Germany. 2011. [30] Hintemann, R.; Fichter, K.: Materialbestand der Rechenzentren in Deutschland – Eine Bestandsaufnahme zur Ermittlung von Ressourcen- und Energieeinsatz, Study for the Umweltbundesamt Berlin, 2010.

Appendix A Evaluation of the innovation potential of selected Low Exergy Solutions The innovation potential analysis (IPA) methodology was developed within the project nordwest2050 (see above). The evaluation phase includes an assessment based on several criteria and was performed on all technologies or combinations of technologies from section 4.3. The criteria include indicators for innovation type, climate adaptation, resilience and sustainability, feasibility, and regional transferability, Here we present the results for the above described pilot project technologies in graphical form.

Figure 1. IPA results for Low Exergy Solution “CHP offheat from biogas plant – absorption chiller – barn cooling” (pilot project 1: “Cooling Turkey Barns”). The further outside the data point, the higher the rating in this category.

162

Figure 2. IPA results for Low Exergy Solution “Geothermal Cooling” (pilot project 2: “Geothermal cooling of a data center”). The further outside the data point, the higher the rating in this category.

Figure 3. IPA results for Low Exergy Solution “Rumen based Biogas Plant” (pilot project 3: “Use of cellulose-rich substrate in biogas plants”). The further outside the data point, the higher the rating in this category.

163

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

On the use of MPT to derive optimal RES electricity generation mixes Paula Ferreiraa , Jorge Cunhab a b

Centre for Industrial and Technology Management, Guimarães, Portugal, [email protected] Centre for Industrial and Technology Management, Guimarães, Portugal, [email protected]

Abstract: The us e of modern portfolio theory (MP T) is a common practice to derive efficient frontiers and support portfolio decision making in financial markets. Although real projects present different characteristics and technical restrictions, the general objective of the decision maker is the same: to maximize the expected return minimizing the portfolio risk. Long term electricity generation decision making is characterized by high uncertainty, high impact on social welfare and a large set of diversified technologies that may be included in future scenarios. The possibility of applying MP T approach to define efficient electricity generation portfolios is explored in this paper focusing on particular in renewable energy sourc es (RES technologies). The us e of MPT for building RES scenarios is demonstrated for the particular case of P ortugal. One year hourly data concerning power output from wind, hydro and solar plants along wit h the power demand was collected and included in t he analysis. Three different approaches were c onsidered for designing the efficient frontiers aiming at maximizing the RES electricity generation, minimizing deviation between the demand and the RES production and minimizing the levelised cost of the RES system. The results demonstrate how t his approach can be an effective tool to support decision making but put also in evidence the need to build modified MP T models in order to take into account the technical restrictions of the system.

Keywords: Renewable electricity sources, Electricity generation, Modern portfolio theory.

1. Introduction Electricity power planning relates to generation, transmission and distribution systems. In this paper we are focused only on the generation system. The main goal of generation planning is to meet customers’ electricity needs at least cost with an acceptable degree of safety, reliability and quality [1]. However, this is a difficult task given that generation planning deals with future decisions that have to be made in an environment of uncertainty (namely, due to electricity demand, fuel prices volatility, investment costs, regulatory framework) and such uncertainties have to be, explicitly, taken into account in electricity planning [2]. In order to achieve this goal, it is necessary to couple supply-side management programs (which involve the construction of new power plants and/or repowering of existing ones) with demand-side management programs (in order to manage the customer load demand) [3]. In short, generation planning tasks include energy and demand forecasting, supply-side management and demand-side management adjustments, analysis of alternative expansion plans, determination of the optimal strategy or portfolio strategies and the evaluation of financial implications and feasibility [1]. Traditionally, the least-cost approach has been used in generation planning. This approach is frequently based on calculating the levelised costs of electricity generation, expressed in €/MWh, for different alternative technologies (e.g. fossil fuels, nuclear, renewable) and comparing such costs in order to choose the technology with the lowest cost. 164

However, some criticisms to the use of this approach can be found in the literature. Firstly, the fact that electricity planning decision makers are faced both with a wider range of alternative technologies for electricity generation and different institutional framework in which they operate, coupled with a future that appears increasingly complex and uncertain [4]. Secondly, as energy markets have been liberalised, the interest in quantifying and manage market risks grew [5]. In fact, with the deregulation and liberalisation of electricity markets, with a corresponding increase in competition, electricity generation companies will no longer have a guaranteed return because the price of electricity varies depending on a number of factors. In this context, it is essential that those companies can manage electricity price risk [6]. Additionally, there is the issue of security of energy supply [7]. In fact, given the global shortage in terms of primary fuel sources, policy makers increasingly need to consider a diversification of electricity production. Simultaneously, the price volatility of fossil fuels raises the question of what are the best options in terms of energy needs of a country. Finally, an important feature of renewable technologies is that they correspond to capital intensive investments, which translates into a relatively fixed cost structure over time, with very low (or practically zero) marginal costs, and that are uncorrelated with important risk drivers, such as fossil fuel prices [6,7]. Given these reasons, it is necessary to shift from a paradigm that seeks to evaluate different technologies for electricity production on a stand-alone basis, to one that evaluate different portfolios of technologies for electricity production [4,7]. This means abandoning the traditional least-cost approach and to adopt a new perspective of analysis based on the theory of efficient portfolios. In this context, the "mean-variance portfolio (MVP) theory is highly suited to the problem of planning and evaluating a nation’s electricity portfolio and strategies" [4]. Although, "at any given time, some alternatives in the portfolio may have higher costs while others have lower costs, yet over time, the astute combination of resources serves to minimize overall expected generating cost relative to the expected risk" [4]. In the context of electricity planning, where a combination of conventional technologies and renewable technologies is being considered, although renewables may present a higher levelised cost, it does not necessarily mean that the overall cost of the portfolio of technologies become more expensive, given the "statistical independence of renewables costs, which do not correlate (or covary) with fossil price movements" [4]. In fact, the inclusion of renewable technologies in an electricity generation portfolio is a way to reduce the cost and risk of the portfolio, although in a stand-alone basis the cost of those renewable technologies might be higher [7]. The electricity generation sector is essential for the attainment of the European renewable objectives. According to the European Union (EU) forecasts, the large hydropower will maintain its dominant position in renewable energy sources (RES) for electricity generation for the near future. However, the use of wind will continue expanding and, in 2020, the onshore and offshore wind electricity generation will overcome the hydro sector in the EU-27. Biomass/waste remains as the third RES for electricity (RES-E) technology with two digit RES share. An increase of the solar technologies is also foreseen although staying far from the wind, hydro or biomass shares [8]. The definition of optimal scenarios for RES-E to include on the grid has been frequently debated in the literature adopting multicriteria tools or electricity planning models based on cost/emissions optimization procedures. However, more recently the importance of diverse electricity technologies portfolios has been also emphasised and the use of the modern portfolio theory (MPT), previously established for the financial investment analysis, has been well applied to the electricity generation sector. This paper applies MPT as an electricity generation planning tool, in order to present optimal RES electricity generation mixes for the future, taking into account the past production pattern of each RES and optimizing the trade-off between maximizing RES output and minimizing RES variability. 165

The rest of the paper is organised as follows. In section 2, a brief description of the MPT reasoning and its application to electricity planning is presented. Section 3 corresponds to the empirical study undertaken, regarding the optimal RES electricity portfolios in Portugal. Finally, Section 4 draws the main conclusions of this paper and presents perspectives for future research.

2. Modern Portfolio Theory for energy decisions 2.1. Brief overview of MPT theory Modern portfolio theory has its roots in the seminal paper by [9]. He proposed a methodology to select efficient investment portfolios based on investors’ goal of maximising future expected return given a certain level of risk they were willing to take [10]. Investors in financial assets expect to earn a certain return over a given investment horizon. However, the yield actually obtained by the investor may differ from the expected return, and this represents the investment’s source of risk. When deciding about his investments, the investor should consider, besides expected return, the following elements [11]: the dispersion of returns around the average return (variance), the symmetry of the distribution (skewness), and the kurtosis of the distribution. However, one of the innovations of the mean- variance model of [9], was the assumption that the distribution of returns follows a normal distribution. This has the advantage of being able to ignore those last two elements because the normal distribution is symmetric and has a kurtosis of zero. Thus, the characteristics of these investments can be measured based on only two variables: expected return and variance [11]. Therefore, assuming the assumption that investors are risk averse, having to choose between two investments with the same standard deviation but different expected returns, they always choose the one with higher expected return (and vice versa). Thus, the mean- variance model allowed to explain the advantages that an investor has to diversify their investments among several securities (e.g. stocks or bonds). That is, instead of investing in a single asset, investing in portfolios made up of various financial assets. In fact, there are two reasons why diversification reduces the risk of investment [11]. On the one hand, as each asset included in a diversified portfolio represents a small portion of the investor’s total investment, any event affecting one or a few of these assets have a more limited impact on the total value of the portfolio. On the other hand, the effect of specific events on the price of each asset included in a portfolio can be positive or negative. In large and well diversified portfolios, these effects tend to offset each other without affecting significantly the overall value of the portfolio. One can illustrate the effects of diversification on the risk of a portfolio by examining the effect of adding more assets to the portfolio and see what happens to its variance. For example, in the case of a portfolio, P, consisting of two assets, A and B, expected return, ( rP ) , and variance, 2 P , are given by, respectively: ( rP ) = A ( rA ) + B ( rB ) (1) and 2 P

= 2 A2 2 A + B2 2 B + 2 A B AB A B (2) where A and B represent the proportions invested in each asset, A and B. The last term in the expression of the variance is often written in terms of the covariance of returns between two assets: AB = AB A B. One can see that the benefits of diversification are a function of the correlation coefficient. Thus, the lower the correlation of returns between two assets the higher the gains from diversification an investor obtain. This reasoning can be generalised for the case of a portfolio with N assets. Thus, expected return, ( rP ) , and variance, 2 P , of the portfolio are given by: i N

( rP )

i i 1

( ri )

166

(3) and i N j N 2 P

i

j

ij

i

(4)

j

i 1 j 1

We conclude, therefore, that the variance of a portfolio is partially determined by the variance of individual assets and partly by the way they move together. The latter is measured statistically by the coefficient of correlation or the covariance of the assets belonging to the portfolio. It is the term for the covariance that provides an explanation of why and in what amount diversification reduces the risk of investment. In fact, portfolios of financial assets should not be chosen only by their individual characteristics, but taking into account how the correlation between assets affects the overall risk of a portfolio [11]. Therefore, since the variances can be estimated for portfolios consisting of a large number of assets, suggests an approach to the optimal selection of portfolios in which investors make the balance between expected return and risk. Alternative 1: If an investor can specify the maximum risk he is willing to take, the optimal portfolio is obtained maximising expected return subject to that risk level, i.e.:

Alternative 2: If an investor specifies his desired level of expected return, the optimal portfolio is the one that minimizes the variance subject to that level of return: i N j N

i N

Max ( rP )

i E (ri )

2 P

Min

i

i 1

i 1

s.t.

ij

s.a. i N j N

2 P

i

j

ij

i N

ˆ2

( rP )

i 1 j 1

i

E (ri )

E( r )

i 1 N

N i

1

1

i i 1

i 1 i

j

j 1

0

i

0

The portfolios that result from this process give rise to what is called the efficient frontier, as represented in Figure 1:

Expe cted retur n

Efficient frontier

Each point on this effic ient frontier represents a portfolio, ie, a portfolio that has the highest expected return for a given level of risk.

Standard deviation

Fig. 1. – Efficient frontier

167

2.2. MPT applications to electricity generation In recent years there has been a growing application of the MPT theory to electricity planning. In fact, the mean-variance model can be used to determine the optimal portfolios of electricity generation both for a company or a country. According to [5], the main idea of the MPT model is that the value of each asset can only be determined taken into account portfolios of alternative assets. Hence, energy planning should be focused more on developing efficient production portfolios and less on finding the alternative with the lowest production cost [4,7]. The MPT approach allows to analyse the impact of the inclusion of renewable technologies in the mix of generating sources of electricity. In particular, it provides a better risk assessment of alternative generation technologies, something that the traditional stand-alone least cost approach cannot do, particularly in terms of the impact of renewable energy sources in reducing the risk of the portfolio of technologies to be adopted. In fact, the MPT model allows to illustrate the trade-off between production costs and risk: the lower the cost the higher the risk, meaning that it is not possible to achieve a lower electricity production cost without assuming higher levels of risk. It should be noted that the result of applying the mean-variance model to generation planning is not identifying a specific portfolio, but the identification of an efficient frontier where the optimal portfolios will be located. These are Pareto-optimal, that is, an increase in returns (or a decrease in costs) is only achieved by accepting an increased risk. On the other hand, an important aspect in the mean-variance model is the assumption that past events are the best guide for predicting the future. Not to say that unexpected events will not occur, but that the effect of these events is already known from past experience [7]. A study that used the MPT theory to obtain evidence about the best mix of electricity generation in Scotland was that of [12]. Based on the efficient frontier, the authors analysed the portfolios suggested in four scenarios for the electricity generation mix in 2020, seeking to clarify what role renewable technologies can play in setting up those portfolios. The main conclusions reached by those authors were that: the portfolios of electricity production corresponding to the four scenarios are not mean-variance efficient; based on MPT approach it is possible to quantify the likely scale of inefficiency; and it seems there is the opportunity to have an improvement in the generation mix in the sense of Pareto. Another study was conducted by [13], where they tried to optimise wind power investment portfolios across countries taking into account the correlation between wind farms output located in different geographical areas. In fact, the aim is "to demonstrate the use of MVP theory as an insightful analytical approach to take into account the impact of wind output variability and correlations of wind output across different locations within a wind farm portfolio" [13]. These authors concluded that the current and projected portfolios for 2020 are far from the efficient frontier and, therefore, there is scope for wider benefits arising from greater coordination of European renewable development by providing "incentives for location of new wind farms so as to maximise the efficiency of the overall European wind portfolio". In turn, [5] apply the MPT theory in order to optimise generation electricity portfolios but focusing their attention "on private investors' investment incentives in liberalized electricity markets, where fuel- mix diversification is a possible strategy for reducing exposure to electricity, fuel, and carbon price risks". In fact, according to these authors, the electric utilities operating in deregulated markets cannot easily pass on to the sales price changes in their production costs. Thus, utilities have to take into account the risks that may affect their profits when they have to decide about its investment projects. In this context, the risks regarding electricity, fuel and carbon prices become relevant in determining the optimal production portfolios. The results obtained by [5] demonstrated the importance of the degree of correlation between the prices of electricity, fuel and carbon in the 168

definition of the optimal generation mix. Hence, they concluded that "liberalized electricity markets characterized by strong correlation between electricity and gas prices […] are unlikely to reward fuel mix diversification sufficiently to make private investors' choices align with the socially optimal fuel- mix, unless investors can find counterparties with complementary risk profiles to sign long-term power purchase agreements". Also from the perspective of a private generation company, operating in a liberalised electricity market, [6] applied the theory of efficient portfolios. In this type of markets, it is essential that utilities companies can properly manage the electricity price risk, given the strong competition among the different operators in those markets. To address this issue, [6] adopt the MPT approach in order to define the best strategy for electricity trading for a company that is considering selling in the spot market or establish bilateral contracts. The question that arises is "how to allocate energy among these potential transactions in order to maximize profits with relatively low risk" [6]. In fact, the combination of different trading strategies of electricity can be seen as constituting a portfolio which can be optimised using the MPT approach. Finally, [4] presents a summary of the application of MPT theory in the evaluation of different electricity generation planning scenarios for the case of U.S., EU and Mexico, where was perceived that the mix of electricity generation can be improved in terms of cost and/or risk, by expanding the use of renewable technologies. The author states that "compared to existing, fossil-dominated mixes, efficient portfolios reduce generating cost while including greater renewables shares in the mix thereby enhancing energy security. Though counterintuitive, the idea that adding more costly renewables can actually reduce portfolio- generating cost is consistent with basic finance theory". It follows an important conclusion: "in dynamic and uncertain environments, the relative value of generating technologies must be determined not by evaluating alternative resources, but by evaluating alternative resource portfolios" [4]. The above mentioned papers demonstrate the possibility of adapting a financial theory on electricity planning problems. In fact, the increase of RES to electricity generation creates important challenges to grid managers due to the expected variability of the power output of most of these RES power plants. The adoption of a model based on MPT can be particularly useful for electricity systems highly RES supported, allowing to take into account both yearly seasonality and intra-daily variations of the production. This paper proposes to demonstrate the use of MPT on these systems resourcing to the particular case of the Portuguese electricity system to identify optimal RES portfolios. The aim is to optimize the trade-off between the variable production that characterize some of the RES and the return of these projects, measured according to a set of proxy variables.

3. Optimal RES electricity portfolios The Portuguese electricity system is mainly based on a mix of thermal, hydro and wind power technologies. RES power plants represent 54% of the total installed power. The wind sector grew rapidly in the last years and an increase on the hydropower investment is also foreseen for the next years, strongly justified by the need to compensate the variable output of wind power plants. As in the EU-27, biomass represents an important RES contributor, mainly because of industrial wastes used in CHP and, in much smaller amount, by the centralized biomass power plants [14]. Some recent studies already addressed the case of electricity generation scenarios in Portugal and the use of optimization models to draw these scenarios [15,16]. However, to the authors’ best knowledge no attempt has been made to use an approach close to the MPT theory to this system. In fact, most optimization models rely on the cost and/or emissions minimization of the electricity system. Functions such as the loss of load probability or the reserve margin are used to address the minimum requirements for security of supply. These functions although allowing to include the variability of RES power output do not explicitly recognize portfolio risk as a decision variable 169

influenced by the risk of each technology output and, most importantly, by the correlations between those risks. The general idea of this research is to present possible RES generation mixes that would ensure maximum return (or minimum cost) for each given portfolio risk level, obtaining then the efficient frontier. The use of the Portuguese case, as an electricity system strongly influenced by RES seasonality behaviour, is expected to contribute to demonstrate how MPT can provide a way to complement cost optimization models with a quantitative risk evaluation of the electricity generation portfolio. The data used for the models was drawn from public information available on REN site (www.ren.pt), consisting of the load output of each RES power plant measured for each quarter of an hour for an one year period. For the case presented in this paper, 2010 information was considered representing 35040 measures for each technology. This allowed to capture the daily and yearly seasonality of RES technologies output and of the demand. Figures 2 to 5 show the load output of wind, small hydro, photovoltaic and small thermal power plants (including renewable and non-renewable cogeneration and biomass power plants).

Fig. 2. Wind power load, Portugal 2010.

Fig. 3. Small hydro load, Portugal, 2010

170

Fig. 4. Photovoltaic load, Portugal, 2010.

Fig. 5. Small thermal power load, Portugal, 2010

From the figures it became evident that the variability of the RES output comes mainly from the non-storage RES production, namely wind, hydro and photovoltaic power plants. The Portuguese system includes also large dams and run of river hydro plants, each one of them with some storage capacity. Although storage capacity of run of river power plants is limited, it also allows reducing the variability of the hydro power output. As for the small hydro power plants most of them do not present storage capacity and as so it was assumed that their production could represent a proxy variable for the hydro availability. Both the wind power and photovoltaic loads were assumed as proxy variables for the underlying resource availability. Being possible to storage, the variability of the biomass power output is much lower than the all the other RES and does not depend on the hourly availability of the resource. For this reason, only, wind, hydro and sun technologies are included in this analysis. To make the variables comparable, the output of each technology was normalized by the installed power in 2010, as described in (5). (5) Where i represents the technology (1- wind; 2- hydro; 3- photovoltaic), t represents the moment in time and Li,t represents the normalized variable for each technology in each quarter of an hour. The demand was also used on the second model proposed, aiming to find the best RES solution that could meet the desired demand with the lowest deviation. For this an additional proxy variable was used to normalise the demand by the peak load, as described in (6). 171

(6) Where LDi,t represents the normalized demand in each quarter of an hour. The proxy variables included on the proposed MPT models are characterized in Table 1 and include: -

Normalized wind power output, representing the wind availability of the system. Normalized small hydro output, representing the hydro inflows (hydro availability) to the system. Normalized photovoltaic output, representing the sun availability of the system. Normalized demand, representing the electricity needs of the system

Table 1. Characteristics of the proxy variables for MPT model. Mean (MW/Installed MW) Standard deviation (MW/Installed MW) Correlation coefficient Wind Hydro Photovoltaic Demand

Wind 0,278 0,210

Hydro 0,383 0,281

Photovoltaic 0,194 0,264

Demand 0,634 0,120

1

0,335 1

-0,255 -0,152 1

0,0019 0,0105 0,0080 1

In the following sections different scenarios will be presented applying models based on the MPT theory. Three different approaches were considered for designing the efficient frontiers: (1) maximizing the RES-E generation (MPT_RES); (2) minimizing the difference between demand and RES-E production (MPT_RES@Demand); (3) minimizing RES cost scenarios, according to the expected levelized cost of each technology (MPT_RES@Cost). Optimization models were built and Excel Solver was used to find optimal solutions for each problem.

3.1. MPT_RES model For this analysis a traditional MPT model was used aiming to design the efficient frontier that can maximize the expected RES production per unit of installed capacity for each risk level. The optimisation model is described by (7) to (10). Objective function

Max Restrictions

(7)

(8)

(9) (10)

172

Where E(Lp ) represents expected return of the portfolio (RES generation per installed MW), Wi represents the share of technology i, E(Li) represents the expected i technology output (i generation per installed MW), (Lp ) represents the standard deviation of the portfolio, i represents the standard deviation of i technology output, and ik represents the correlation coefficient between i and k technologies outputs. Figure 6 and Table 2 describe the results obtained, including the efficient frontier and the characterization of a set of optimal portfolios. Figure 6 presents also the present RES (wind, hydro and photovoltaic) portfolio and the expected one in 2022, according to REN forecast [15,16].

Fig. 6. Efficient frontier MPT_RES model Table 2. Characterization of MPT_RES optimal portfolios Portfolio 1 Portfolio 2 Portfolio 3 Portfolio 4 Portfolio 5 Portfolio 6 2010 Scenario 2022 Scenario

(Lp ) 0,28 0,25 0,22 0,2 0,18 0,15 0,194 0,18

E(Lp ) 0,383 0,369 0,354 0,341 0,327 0,299 0,336 0,327

Wind 0,30% 13,00% 27,82% 34,58% 36,54% 40,68% 42,03% 38,22%

Hydro 99,70% 87,00% 72,18% 62,50% 54,13% 37,52% 56,59% 53,46%

Photovoltaic 0% 0% 0% 2,92% 9,33% 21,80% 1,38% 8,32%

3.2. MPT_RES@Demand model For this analysis a modified MPT model was used aiming to design the efficient frontier that can minimise the deviation between the demand and the RES production in each moment. The idea is to define optimal RES portfolios that can contribute to better meet the demand in each moment, following a close load distribution pattern. The proposed optimisation model is described by (11) to (14). Objective function Min (11) 173

Restrictions

(12)

(13) (14) Where, d represents the standard deviation of the demand and id represents the correlation coefficient between i k technologies outputs and the demand. From the reduction of risk perspective, a negative correlation between technologies is desirable to ensure their complementarity. However, but a positive correlation between RES technologies output and the demand should lead also to risk reduction under this model. The traditional standard deviation calculation was changed taking this into consideration, as may be seen in (12). Figure 7 and Table 3 describe the results obtained, including the efficient frontier and the characterization of a set of optimal portfolios.

Fig. 7. Efficient frontier MPT_RES@Demand model Table 3. Characterization of MPT_RES@Demand optimal portfolios Portfolio 1 Portfolio 2 Portfolio 3 Portfolio 4 Portfolio 5 201 Scenario 2022 Scenario

(Lp ) 0,304 0,28 0,25 0,22 0,20 0,227 0,215

E(Lp ) 0,25 0,262 0,279 0,302 0,322 0,297 0,306

Wind 0% 11,48% 27,87% 35,87% 38,76% 42,03% 38,22%

174

Hydro 100% 88,52% 72,13% 56,40% 44,52% 56,59% 53,46%

Photovoltaic 0% 0% 0% 7,73% 16,72% 1,38% 8,32%

3.3. MPT_RES@Cost model This analysis is similar to the one conducted in section 3.1. However, the model is now weighted by the levelised costs of each RES technology. This way, an efficient frontier will be drawn from the optimization model with the objective goal being the minimization of the total expected cost of the RES system. The optimization model is described by (15) to (18) describe the model. Objective function

Min Restrictions

(15)

(16)

(17) (18) Where E(LCp ) represents the expected levelised cost of the portfolio per unit of installed capacity, (LCp ) represents the standard deviation of levelised cost of the portfolio and LC i represents the levelised cost of each i technology . Figure 8 and Table 4 describe the results obtained, including the efficient frontier and the characterization of a set of optimal portfolios.

Fig. 8. Efficient frontier MPT_RES@Cost model 175

Table 4. Characterization of MPT_RES@Cost optimal portfolios Portfolio 1 Portfolio 2 Portfolio 3 Portfolio 4 Portfolio 5 Portfolio 6 Portfolio 7 Portfolio 8 2010 Scenario 2022 Scenario

(LCp ) 18 16 15,5 15,17 14,5 14 13,5 13,3 16 15,53

E(LCp ) 29 27,51 26,8 20,13 21 21,77 22,75 23,29 27,92 29,71

Wind 31,80% 44,84% 48,80% 100% 94,36% 89,90% 84,12% 81,03% 42,03% 38,22%

Hydro 68,20% 55,16% 51,20% 0 5,64% 8,28% 12,74% 15,10% 56,59% 53,46%

Photovoltaic 0 0 0 0 1,82% 3,14% 3,87% 1,38% 8,32%

3.4. Analysis of results The results indicate that both 2010 and 2022 scenarios [17,18] are close to the efficient frontier for MPT_RES and MPT_RES@Demand models. In fact, both these scenarios reflect the Portuguese energy policy goals of increasing RES share on the electricity system, diversifying the energy sources and promoting a strategy based on hydro reinforcement to deal with the increasing wind share. In the same way, most of the less risky scenarios described in figures 6 and 7 point to mix hydro-wind power scenarios as the more efficient ones. More risky strategies rely mainly on hydro power, the option with higher expected return but also the one with higher standard deviation. Although a positive correlation exists between wind and hydro, it does not seem to be enough to jeopardize the mix of these technologies in most of the scenarios. On the other hand, photovoltaic presents a less interesting expected value and a risk level close to the hydro one. It presents, however, the advantage of being negatively correlated to both wind and hydro. As so, less risky scenarios tend to include also this option combined with hydro and wind. The MPT_RES@Cost present quite different results, clearly driven by the levelised cost of the technologies. A strong reliance on wind power is evident along the efficient frontier, as this is the option with less expected cost and with the lowest standard deviation when considering the levelised cost normalized by the installed power. Solutions with lower risk are characterized by a mix of wind, hydro and to a much lower extent photovoltaic technology, leading to a higher expected cost but also taking advantage of the portfolio diversification. Particularly interesting for the MPT_RES@Cost is the comparison of portfolio 3, portfolio 4 and 2022 scenario. All of these solutions have close risk values, but very different expected levelised costs and RES structures. What seems to be the best solution (portfolio 4) is however, compromised by a 100% wind power share. From the technical point of view is a nonsense solution, due to the already existing hydro capacity and for motives of security of supply. Both portfolio 3 and 2022 scenario are much more balanced solutions, as a stronger diversity of the mix is foreseen. The obtained results put in evidence the need to enrich the traditional MPT analysis with additional technical, legal and economic constraints when passing from financial markets to the analysis of portfolios of real projects. Traditional strategic electricity power planning cost optimization models with technical restrictions must be combined with efficient portfolio design with risk restrictions. The research project is now proceeding with this new approach into a single quantitative framework, envisaging the following elements: A cost objective: to minimize levelized cost of production of the electricity system as a whole. An environmental objective: to minimize environmental impacts, either measured by emissions or by externalities valuation. 176

A risk objective: measured by the variance of the portfolio. A set of decision variables: share of each technology, measured by the ratio between the installed power of each technology and the total installed power of the system. A set of constraints: capacity limitations, legal and technical requirements and electricity demand needs.

4. Conclusion Social welfare strongly depends on a reliable and competitive electricity system. RES technologies constitute key investments to design future scenarios or strategies for sustainable future. The raising trend of RES brings however considerable challenges to decision makers due to uncertainty of the production highly dependent on the availability of the underlying resources. This paper demonstrates the application of MPT for RES in electricity planning. This allowed to address both the expected return and the RES portfolio risk, taking into account both the standard deviation of each technology output and the correlation coefficient between technology outputs and demand needs. The study of the Portuguese case concludes that less risky solutions are characterised by a mix of RES technologies. This mix, however, depends on the criteria used to quantify the expected return. If the maximisation of the RES contribution to the system is the goal of the planner, hydro emerges as the major contributor. On the other hand, if decisions are driven by levelised costs, hydro is penalised and wind becomes the preferable option. Photovoltaic share only becomes relevant for low risky solutions, regardless of the model used. The present Portuguese RES generation mix and the forecasted scenario for 2022 [16, 17] showed to be close to the efficient frontier for the case of MPT_RES and MPT_RES@Demand models, reflecting the diversification goal for the sector. Notwithstanding, when the levelised cost is included in the analysis, both 2010 and 2022 scenarios move away from the efficient frontier. Although the usefulness of the MPT approach in analysing the electrical planning scenarios, has been demonstrated, it is important not to forget some limitations of this approach. For example, [12] emphasised two issues. On the one hand, the failure to consider transaction costs associated with changes in generation mix. Second, the fact that, generally, the studies carried out do not take into account the feasibility of the efficient portfolios obtained with the MPT theory in the context of existing energy infrastructure. Moreover, [7] pointed out that the characteristics of electricity generation technologies are not always comparable to the characteristics of financial assets for which the MPT theory was developed. Firstly, markets for assets (e.g. turbines, coal plants) related to electricity generation are usually imperfect in contrast with capital markets, which also make them less liquid. Secondly, financial assets are almost infinitely divisible and fungible, which does not happen with electricity generating real assets. Finally, investments in electricity production technologies tend to be lumpy, especially renewable technologies. However, [7] consider that "for large service territories or for the analysis of national generating portfolios, the lumpiness of individual capacity additions becomes relatively less significant”. Recognizing that MPT for electricity system analysis must go beyond the traditional models, future work envisages the development of a new model combining MPT with generation expansion models for electricity power planning.

Acknowledgments This work was financed by: the QREN – Operational Programme for Competitiveness Factors, the European Union – European Regional Development Fund and National Funds-Portuguese 177

Foundation for Science and Technology, under Project FCOMP-01-0124-FEDER-011377 and Project Pest-OE/EME/UI0252/2011.

References [1] Beltran H., Modern Portfolio Theory Applied to Electricity Generation Planning [Master dissertation]. Illinois, USA: University of Illinois at Urbana-Champaign; 2009. [2] Joode J., Boots, M., Concepts of investment risks and strategies in electricity generation. Energy Research Centre of Netherlands, June 2005, ECN-061. [3] Sedano R., Cowart, R., Power system planning and investment, New England Demand Response Initiative (NEDRI) Report, March 2003. [4] Awerbuch A., Portfolio-Based electricity generation planning: policy implications for renewables and energy security. Mitigation and Adaptation Strategies for Global Change 2006; 11:693–710. [5] Roques F., Newbery D., Nuttall W., Fuel mix diversification incentives in liberalized electricity markets: A Mean–Variance Portfolio theory approach. Energy Economics 2008;30:1831–1849. [6] Liu M., Wu F., Portfolio optimization in electricity markets. Electric Power Systems Researc h 2007;77:1000–1009. [7] Awerbuch S., Berger M., Applying portfolio theory to EU electricity planning and policymaking. IEA Research Paper, Paris, February 2003, Report Number EET/2003/03. [8] European Commission, EU energy trends to 2030- Update 2009; 2010; http://ec.europa.eu/energy/index_en.htm [9] Markowitz H., Portfolio selection. Journal of Finance 1952;7(1):77–91. [10] Huisman R., Mahieu R., Schlichter F., Electricity portfolio management: Optimal peak/offpeak allocations. Energy Economics 2009;31:169–174. [11] Damodaran A., Corporate finance: Theory and practice (2nd ed.). John Wiley & Son; 2001. [12] Allan G., Eromenko I., McGregor P., Swales K., The regional electricity generation mix in Scotland: A portfolio selection approach incorporating marine technologies. Energy Polic y 2011;39:6–22. [13] Roques F.; Hiroux C., Saguan M:, Optimal wind power deployment in Europe — A portfolio approach. Energy Policy 2010;38:3245–3256. [14] DGGE, Renováveis-Estatísticas rápidas; Janeiro 2012; www.dgge.pt (in Portuguese) [15] Kraja G., Neven D., Carvalho M.G., How to achieve a 100% RES electricity supply for Portugal?. Applied Energy 2011; 88(2):508-517. [16] Pereira S., Ferreira P., Vaz A.I., Strategic Electricity Planning Decisions. Proceedings of the 6th Dubrovnik Conference on Sustainable Development of Energy, Water and Environment Systems; 2011 September 25-29; Dubrovnik, Croatia. [17] REN, Dados Técnicos 2010; 2010; www.ren.pt (in Portuguese). [18] REN, Plano de Desenvolvimento e Investimento da Rede de Transporte de Electricidade 2012-2017 (2022); Julho 2011; www.ren.pt (in Portuguese).

178

PROCEEDINGS OF ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

 

Stability and limit cycles in an exergy-based model of population dynamics Enrico Sciubbaa, Federico Zullob   a

Dept. of Mechanical & Aeronautical Engineering, University of Roma La Sapienza, Via Eudossiana 18, 00184 Roma, Italy. [email protected] b

School of Mathematics, Statistics & Actuarial Science, University of Kent, Canterbury CT2 7NF, UK. [email protected]

Abstract The long-term qualitative behaviour of most real ecological populations can be mathematically classified as either asymptotic equilibrium or limit cycling. The reasons underlying these types of behaviour have been, and are still, debated. Aim of this paper is to show that it is possible to characterize both of these behaviours through a population dynamics model, proposed and developed in previous work by the authors, in which the resource consumption is quantified solely in terms of exergy flows. A proper number of reasonable assumptions on the phenomenological characteristics of the interacting species result in different evolutionary scenarios, corresponding respectively to asymptotic stability or periodic attractors. In the case of asymptotic periodic motion the system is described by a set of non-linear, non-autonomous differential equations and an analytical investigation becomes arduous even in the simplest cases. Explicit results, both analytical and numerical, are discussed in this paper. The theoretical description is compared with the behaviour of real, well-known biological systems: the snowshoe hare-lynx predator-prey system for the periodic case and the reindeer herds of the Pribilof Islands for the case of asymptotic stability. The assumptions that must be posited to obtain a cyclic limit type behaviour in ecological populations appear to have been never recognized before and the possibility to adopt them as an universal mechanism of population cycles is discussed in the conclusions.

Keywords:

sustainability, eco-systems analysis, extended exergy accounting, population dynamics, population cycles.

1 - Introduction  

The concept of sustainability is inherently linked with the consideration that any analysis, be it qualitative or quantitative, of system-environment interactions unavoidably leads to the identification of a certain number of natural resources, each available in a finite amount. Since the dynamic characterization of the interplay among one or more species in a given ecological niche cannot ignore their dependence on the resources they can avail themselves of, this leads to an “intrinsic” definition of sustainability. If one adopts this point of view, there are some difficulties that preclude the construction of sufficiently detailed mathematical models that may describe, within a general framework, the outcome of the struggle for life. Indeed any model must confront itself with three well-known empirical results: i) the multiple “modes” of natural behavior detected in all existing species, that can be qualitatively categorized into interactions such as competition, cooperation, antagonism, adaptation and so on; ii) the broad variety of resources each species can 179  

   

avail itself of; and iii) the natural “flexible attitude” of most species towards the substitutability of certain resources. As in most modeling activities, and specifically in those attempting to treat living organisms, there exists a fundamental competition between the adoption of a completely heuristic approach, that leads to the development of several and possible unrelated ad hoc models, each of them focusing on the particular dynamics of a certain species, and the scientific urge for generality, that favours higher-level attempts to interpret global processes as obeying a general trend. The first approach bears the risk of describing in a sufficiently predictive way only tiny scraps of the surrounding reality, while the second, to be successful, must somehow overcome the three above stated problems. The point of view expressed in our previous papers [1-2] is that this second approach can be properly formulated by means of an exergy analysis, or, to be more precise, in terms of the concept of extended exergy (EEA, [3]), a physical quantity that allows to objectively measure and compare the different forms in which natural resources make themselves “available”. Use of EEA makes it possible to use a single additive measure for the resources available to populations. With regard to the three points raised above, it is clear that no genuinely universal approach can be mathematically formulated, because this would imply that the analyst is able to classify within general schemes even voluntary or instinctive behaviors that, especially in superior species, must be considered only in relation to a (species-specific) framework of social norms. Nevertheless, if one assumes, following Lotka [4], that consciousness is closely bound with life processes, it becomes possible to posit a somewhat general working hypothesis, formulate a proper mathematical model on its basis, and then infer from the model some global (i.e., at system level) and “local” (i.e., species-specific) consequences that must subsequently be validated by some suitable in vivo experiment. Such an approach has proven successful, as it was demonstrated in two previous papers [1-2], where an analysis of the sustainability concept made on the basis of a competitive model of populations dynamics was presented and discussed. The “competition” was among the species and for the available resources. The hypotheses underlying that analysis are very simple: 1) The number N of individuals in a species at a given time depends (non-linearly) on the global primary exergy resources a population can avail itself of. 2) The global primary exergy resources may be quantified by the Extended Exergy method [3]. The resulting equations, coupling the exergy rate E (t ) of common resources exploited by Z populations, with their respective population sizes, read: Z     ri E (t )   i  aij N j (t )  j 1   N i (t )  N i (t ) Z   E (t )   aij N j (t )  j 1  

1)

where the dot denotes the time derivative, ri , i=1…Z, are the growth rates of each populations for infinitely abundant resources and  i ’s are the intrinsic mortality rate of each species when the availability of resources drops to zero. The diagonal elements a ii of the system matrix can be interpreted as the specific exergy consumption rate corresponding to a “minimal survival” for the population labeled with the index “i ”, while the off-diagonal elements (ij) are a measure of the strength of the competition between populations i and j (ank=akn=0 meaning that the two populations n and k do NOT compete). Once a well-defined scenario is provided by supplying an

180  

   

independent equation for the time history of the exergy rate E (t ) of the exploited resources, equations (1) determine the evolution of the Z populations. For a detailed analysis on the dynamics of competing populations governed by eqs. (1) readers are referred to [1]. In this article, on the basis of the previous assumptions, we analyze some universal behaviors in the dynamics of biological systems: the adjective “universal” implies that they are shared by a large number of case studies on the dynamics of given species populations. Indeed most of the behaviors are either limit cycles or stable equilibrium [5], [6]. As for the population cycles, many hypotheses have been developed in the last fifty years in order to clarify the physical causes underlying their dynamics. Generally speaking, three orders of difficulty are encountered: i) why does the period of the empirically observed cycles change from species to species (and, under apparently similar circumstances, even for the same species)?; ii) how can the amazing synchronic fluctuations of the numerosity of some species in a given geographical region be justified?; and iii) what is the link between a given species undergoing -or not- cycles and the state of its immediate environment? The existing hypotheses can be classified in abiotic (weather and sunspots), biotic intrinsic (genotype or behavioral changes) and biotic extrinsic (food, parasites, predation etc.) [7-8]. Surely some of these hypotheses can explain at least partial elements of the cycles problem: for example the ten to eleven-year sunspot’s cycle could explain the ten-year cycle of hares but not the four or five year cycles of voles [8]. Also, most of the proposed models do not answer all the above three questions, essentially for two reasons: first, it is very unlikely that the answer to these questions rests with one single physical explanation, and second, no model divulged to date can explain all sides of the corresponding experimental evidence. In this paper we propose an answer to all of the three questions, obtained by positing a constructive assumption (in the sense of Popper’s definition [9]): we develop and discuss a model that provides a good qualitative and an approximate quantitative description of population cycles and thus can be tested for experimental validation or falsification. To the best of our knowledge, the assumptions that constitute the scientific basis of our model have never been presented before. As an application of our model, we present a comparison with the well-known empirical observations of the lynx-hare system [10-11]. The second issue that we consider in this paper is the asymptotic stability of the numerosity of some populations: indeed also this case is very common in both in natura and in vivo observations (see for example [6]). Our central assumption is that, since the numerosity of a population depends on the exergy resources it can avail itself of, the cause of this type of behavior must be searched in the term in eq.1. As shown in [1-2], if the evolution of this term turns out to drive the solution to an asymptotically stable point (note that this includes both scenarios with non-renewable resources and scenarios with a constant amount of renewable resources), then the populations will follow this stability and the long time picture can be either sustainable (survival of all species in certain numerical ratios) or non sustainable (there will be an extinction of some or all of the species), depending on the biotic parameters of the species. An explicit case, the reindeer herds of the Pribilof Islands [12], will be discussed in detail, both qualitatively and quantitatively. In passing, it will be shown that the proposed approach can also answer a longstanding problem about the different behaviours of two herds living in very similar niches.

2 – Limit cycles of populations as a result of differentiated exergy sources: a fundamental hypothesis. 181  

   

The hare-lynx interaction is a well-known process on which an extensive body of experimental and theoretical literature exists: therefore we have chosen it as a benchmark for our model. First of all, let us recall some well-known empirical facts, collected in a large number of publications, about population oscillations. In the case of a predator-prey system, the causes that lead to numerical fluctuations of the two species seem to be different. It is known that cycles often occur in areas of low prey diversity, where specialist predators feed primarily on a single prey species [13-15]. Indeed the diversity of mammalian prey decreases from south to north in the northern hemisphere; when there are alternative prey to hares, the lynx diet becomes differentiated. In [15] the authors take into consideration the relationship between the lynx dietary specialization and the cyclicity of their population. They come to the conclusion that an increase in specialization leads to a truly cyclic dynamics of predators, whereas when the predators can vary their diet feeding on alternative prey the dynamics is no longer cyclic. These considerations support the assumption that the lynx “track” the oscillations of the hares rather than causing them with their predatory pressure. This shifts the attention to the cause of the prey fluctuations. In [5], where about 700 time series (covering more than 25-years) of different species are analyzed, it is argued that, for mammalians, the probability to observe oscillations increases with latitude, but there are no such trends for the amplitude or the period of the oscillations. On the other hand it is also observed [16] that environments experiencing strong seasonality turn out to be more favorable to the onset of larger amplitude cycles. Now consider the simplest case: the presence of a prey species in a given environment without predators. If our assumptions are correct and the fluctuations of the hares are not caused by the predatory pressure, it ought to be possible, depending on the type of environment (represented here by the exergy input), to observe oscillations also in this simple situation. Now let us posit that: i) The environment without prey corresponds to a steady dynamic state: the exergy flow, that in presence of prey is partially exploited by the species, is, in the absence of this species, constant. ii) The total exergy flow is given by the sum of different inputs, none of which is essential for the species survival.

Fig. 1: The case of multiple incoming exergy sources available to a species. In fact, the total primary exergy flow (basically, solar irradiation, water and nutrients from the soil and atmosphere/hydrosphere) will be exploited by the (biotic and abiotic) systems in that niche who will reach some steady state dynamic equilibrium (for obvious reasons, we exclude from our considerations situations in which the competition at these “elementary” scales leads to a global extinction). For assumption ii) we are thinking of an environment offering an ensemble of food sources, each of which is consumed randomly by the prey species: if the consumption of a given source is so intensive that it is completely depleted in a given time, then after this time the species will thrive on some or all of the other sources; also, if the source depleted is renewable (like the 182  

   

vegetable ones), it will be in part or in total restored. As we will now show, under this assumptions the prey species will thrive on a periodic total exergy flow. Thus our model assumes that when the preys are absent from their environment, the value of the “secondary” exergy flow (the one that is indeed primary for the prey species) is constant in time, . (= assumption 1). In presence of prey will no longer be constant, so that say there is now an equation defining the variation of around the point . We need a differential equation defining the evolution in time of . It is clear that such evolution cannot be determined solely by the knowledge of the initial condition | , because one has to take into account that can be a decreasing function in a neighborhood of t=0 due to the predominance of the action of the prey with respect to the force tending to restore the resource, or can be an increasing function in the same neighborhood due to the predominance of the external input with respect to the prey consumption. This means that we need also the value of | , and the equation defining the evolution of must be a second order differential equation. So we need of an equation defining . Suppose that the prey consume their food only in an interval of time, say from to ; after the situation will tend to return in equilibrium as before , because of the continuous primary exergy input into the system. To larger fluctuations around the point correspond larger values of for , that is the restoring force: this means that depends on the difference some function F. By assumption 2, we can expect that to a greater variation of a around the point there corresponds a smaller number of exergy sources. Vice versa a large number of exergy sources will correspond to smaller variations: in this case it suffices, in order to extrapolate the dynamic of the sources to expand F around the point and retain only the lower terms in the expansion:

~

⋯.

So that the resulting equation for

2) is given by: 3)

Note that the minus sign on the term is due to the “restoring” nature of the force tending to the re-establishment of the steady flow solution , whereas the term has to be zero because of the steady flow solution for . The time evolution of the variable is then given by:

sin

.

4)

According to assumption 2, the amplitude of the oscillations number of food sources (Fig. 1).

must be a decreasing function of the

Notice that the exergy flow 4) is the flow available to the prey, whether the prey represent the available exergy input for the specialist predator. If in equation 1) one substitutes ( = hares) for , (= lynxes) for and the flows for the corresponding exergy inputs, the following equations are obtained:



5)

0. The constant v has dimensions of J/kg and Where, without loss of generality, we set provides the conversion of the mass of the prey into exergy utilized by the predators. Obviously here 0 because the prey do not compete with the predators (hares are assumed not to feed on 183  

   

each other!). The set of equations 5) is not amenable to an exact treatment: it is a system of nonlinear, non-autonomous equations. So we first try an analytic approach to the simplest of the cases: the presence of prey without predators. One obtains:



6)

Let us discuss some global features of this equation. Since the set 0, 1. . submanifold for equations 1), we consider only the line 0. The curve

is an invariant

(the grey curve in fig.2 ) divides the semi-plane , 0 into two parts; in the upper one the 0), in the lower one the population increases ( 0): after an initial population decreases ( transient, the solution stabilizes and oscillates within a band width , as in the two examples of figure 2. So, the set

,

is the attractor of our system. Notice also in

figure 2 how the solution curves decrease or increase according to their position with respect to the reference grey curve .

Fig.2 After an initial transient where the population presents an exponential growth or decay corresponding to an initial condition 0 0 or 0 , the population stabilizes on a given curve. As we will show this curve is indeed independent from the value of 0 in the sense that for whatever initial condition, all solutions converge on the same curve. This is clear also by inspecting fig. 2, where the evolution history of the prey population for three different initial conditions is shown. Without loss of generality then, rather than the initial transient, we are interested in this long-term, oscillating behavior of the solution, which is independent on initial conditions. The extrapolation of 0, the set this behavior can be suitably done noticing the role played by the parameter . When ∗ reduces to the point : in this case indeed there is no oscillation , ∗ . in the exergy sources and the species reaches asymptotically its carrying capacity given by When is not equal to 0, the solution oscillates within a set whose amplitude is of order , so the oscillations will be at most of order . This suggests to expand the dynamical variable in

184  

   

series of the dimensionless parameter

≡∑

. After inserting the ansatz , one finds that the functions

6) and equating the same powers of

in eq. are solutions of the

system:

0 sin 0

sin … …



sin

0

2 7)

This system can be, in principle, solved iteratively. In fact the first equation involves only the unknown function . When solved, one can insert the result into the second equation and find and iterate. Note also that only the first equation is non-linear (but indeed easily solvable), the others are linear differential equations of first order in the highest derivative, so the iteration can be carried out by using standard techniques. A further simplification in the calculations can be obtained by noticing that, if the only asymptotic solution (the oscillating one) is sought, it is possible to find the asymptotic behavior of and then solve the equation for . The terms containing the information about initial condition 0 will always decay exponentially in time, so in the asymptotic regime we can write:

sin

5

cos 2

2

2 2

where the convenient parameter

cos

4 … …

4

sin 2

has been introduced. In the fig. 3 we report a plot of two

of the system 6) corresponding to two different initial conditions (grey numerical solutions curves), the reference curve (dotted grey curve), the solution until the first order in the oscillating regime (

) (dashed black curve) and the solution until the second

order in the oscillating regime (

) (dotted black curve). As shown in the

enlarged detail on the right, already at the second order the convergence of the series is very good, since the numerical solutions (marked in light-blue in the zoom) and the approximated one practically coincide.

185  

   

Fig. 3

We see by inspection that the mean value of is affected by the oscillations only at the second order. Indeed if there are no oscillations in the exergy flow the prey would reach the carrying . In presence of oscillations the population is approximated by capacity limit given by ∗ and the arithmetic mean on a period of oscillations

is given

by:

1

that is a bit lower than

∗

∗

. Since

and

1

2

are positive, the function

is bounded,

and more precisely:

0

2

1 2

So that at second order, the difference between

and

∗

is at most

. Therefore if the

mean exergy flow to our population is , the mean of the numerosity of the population is not ∗ as it would be expected, but it is a bit lower. given by the value Let us now introduce the predators into the picture (eq. 5)). In order to describe analytically the solution we try the approach of the series solution as before. We look at the invariant submanifold and exponentially 0, 0 . Also in this case it can be shown that the two curves converge to a periodic curve with the same period of the incoming exergy and that this asymptotic behavior is again independent on the initial conditions, in the sense that whatever initial condition will push the solution on the same curve . For a more detailed discussion of the role played by these type of solutions in living systems we refer readers to [17] and references therein; here we wish 186  

   

only to underline that, being independent from the initial condition and globally attracting [17], these curves and their mathematical description represent the nearest description of the notion of dynamical equilibrium developed (and observed) in ecological systems.

≡∑

Inserting the series

equating the same powers of

and

≡∑

into equation 5) and

one obtains again a linearization of the system 5), in the sense

that the two terms of order 0, that is and solve a coupled non-linear system with constant , , 0, solve a recursive linear system of differential coefficients, while all the pairs , ,0 . Again a further simplification can be equation, depending on all the functions obtained by taking, in the recursive process, only the asymptotic part of each term. For the first couple of equations ( 0) one obtains: 8) The second couple of equations, that is the equations for , , shows that both linear combinations of sin and cos , so also in this case the mean values of



on a period are given by they differ from

and

1

only at order

and



. Note also that the value of

and and

1

are

, that is

is proportional to the

growth rate of the predators but also to the growth rate of the prey. Vice versa the value the prey is proportional to their growth rate and to the mortality rate of the predators.

of

3-Comparisons with field data. In this section w briefly discuss a comparison with field data. The time series for the lynx-hares is well-known (see for example [10-11]), so here we refer to this set of data, ranging from 1845 to 1935. To establish a working methodology we assume that the series are indeed periodic: with this assumption we mean that there exists some curve-fitting algorithm, for example the least square method, giving a set of parameters that specify which periodic function is the best approximation of the data. However we can always write the periodic function as a series in sine and cosine, so we would expect, for the function and something like:

∑ ∑

cos cos

sin sin

9)

For example at first order it can be calculated that the best least square fit for the data of the hares is = 44.81+13.96sin(0.64t)+24.91cos(16/25t), where t is expressed in years and H(t) is given by expressed in thousands. The series 9) is the same type of the series that can be obtained, at least for reasonably low values of n, solving the system 5) by series expansion as explained in the previous section:

∑ ∑

cos cos

sin sin

10) 187

 

   

The coefficients , j=1,2 and will depend on the constants appearing in the system , , 5), so the matching among the coefficients of 9) and 10) will give some constraints on the constants. If one is satisfied with the first order, than there is always a perfect matching between 9) , and 10) because we have an underdetermined system, that is six equations ( , ) in ten unknowns that are the constants appearing in 5) (or, better , , , nine unknowns because we are always free to rescale some set of constants and the system remains unchanged). The second order is more difficult to manage (and indeed we have not succeeded to represent it appropriately) because the coefficients , , j=1,2 depend in a very complicate manner by the constants appearing in 5) and the corresponding system of equivalence among polynomial of high order is not amenable to a close form solution even with the aid of symbolic manipulation software, This case appears though to be interesting because we have ten equations in nine unknowns.

4. Asymptotic equilibrium: the case of the reindeers of the Pribilof Islands. The history of the reindeers herds placed by the United States on the Pribilof Islands in 1911 is well known in literature (see for example [12]). Let us recall only the essentials points of the story for what concerns our purposes. The main islands of the Pribilof group are the St. Paul island (extension about 100 km2) and St. George Island (about 90 km2). In 1911 25 reindeers were placed by the U.S. government on St. Paul and 15 on St. George. On the islands the reindeers did not have any predator (except for a very limited hunting by men). By the very beginning these conditions attracted the interests of ecologists because of the possibility to use the herds as a model to study the introduction of other groups on a larger scale. From this point of view the time-history of the herds can be considered as an “in natura” experiment. The lichens vegetations of the islands seems to have a large influence on the behavior of the herds. As pointed out by Palmer [18], in the winter the reindeers prefer to feed on lichens: from December to April their desirable forage consists of 75%90% of lichens, mosses and other vegetations. Palmer concludes: “although the lichens can not be said to be necessary for the reindeer maintenance because of their nature or nutritive qualities, yet from the standpoint of a readily accessible winter food supply they are essential” [18]. According to this author, in an environment similar to those of the Pribilof Islands, the lichens take from 15 to 20 years to recover a grazed area; other authors (see for example Hanna [19]), based on direct inspections, lean towards much shorter period, such as 5 or 6 years. It is clear however that a continuative consumption of lichens by reindeers tends to make this resource non-renewable. As witnessed by Scheffer in 1951: the food lichens are now so rare […] that diligent search is required to find representative specimens [12]. So we can conjecture that, as concerns our model 1), we have one populations, the reindeers, thriving on non-renewable and renewable resources. The renewable part comes from the process of re-growing of that part of grazed food vegetables having a higher growth rate and of the lichens and mosses themselves. Indeed if the number of reindeers were not so large to consume all the food in a certain time then a carrying capacity would be reached, and in our model this carrying capacity limit is always associated to a renewable exergy flow exploited by the population [1-2]. Here we see also the mathematical characterization of equations 1) in the case of only one population and an exergy flow given by a mix of non-renewable and constant renewable resources. In the rest of this section we try to compare the analytical and numerical results with the time series of the reindeers herds on St. Pauls and St. George [12]. Let us report, for the sake of completeness, the equations describing the system:

188  

   

  rE (t )  cN (t )    N (t )  N (t )   E (t )  cN (t )     E (t )  E (t )  E  bN (t ) E (t )1   E nr (t )    E nr r nr r    E    M  

11)

We must remark that in general equation 11) are the scaled version (in time) of similar equations, where the functions N and E are premultiplied by a constant that, however, can be absorbed in the definition of time. In these formulae N(t) indicates the numerosity of the population, is the total exergy flow incoming the system, and being respectively the non-renewable and renewable part of this exergy, is the total amount of cumulative exergy available from the non-renewable resources, r is the intrinsic or genetic growth rate of the species, is the mortality rate for the population in the limit of vanishing resources, c is a parameter giving the specific exergy consumption rate corresponding to a “minimum survival” of the species, b is a scaling parameter indicating how fast the non-renewable part of the resources are exhausted. With intrinsic or genetic growth rate we mean that obtained as the difference between the birth rate and the mortality rate of the species when no limit is set to the availability of the resource. As a lower bound for this value we can obviously take the maximum observed growth rate for our population ( 2.0/ in our case), whether as a lower bound for we can take the maximum of the mortality rate for our in our case). The true values can differ from these (they can be a little population (~0.9/ higher) since a more realistic inspection shows that the “best possible” real environment for the species is both different from the “infinite” resources scenario and the “worst” one somewhat above the “zero” resources scenario. For the value of c we have an estimate from [20] were a value of 1,735 kcal/days for non-pregnant females and 2,829 kcal/days for pregnant females is given as the mean daily energy requirement between April and May. During the winter the values can be higher, however the order of magnitude remains the same. Reported in Joule and years we have a value c ~ 4·109 J/years. To obtain an order of magnitude for the value of we can assume that the reindeers (or the equivalent form thrived only on non-renewable resources. Then the value of c∑ for the discrete case), where is the time when the population goes extinct, gives a lower . bound for . For the St. Paul island we have calculated a value ~ 9·1013 J, whereas for the 13 . St. George island we have a value of ~ 1·10 J. The value of b is more difficult to quantify, because it can depend, other than on the particular species under consideration, also on several environmental factors. For example a crust of ice in the winter can hinder the grazing of the reindeers. So we have carefully changed the value of this parameter to fit the experimental data. The results are shown in the following figures.

Fig.4: the time series of the reindeers herds on the Pribilof Islands. 189  

   

The two plots on the left show the data on the populations from [12] (crosses) and the curve obtained by equations 11). The values of the constants are as follows: for the St. Paul island we set r=3/years, =2/years, R=4·109J/years, c=4·109J/years, =9·1013J, b=1.35·10-2/years, 10 whereas for the St. George island we set r=3/years, =2/years, R=4·10 J/years, c=4·109J/years, =1·1013J, b=1.25·10 -1. In the plot on the far right we report the two populations on the same axes in order to provide a graphical representation of the very different behavior of these two herds. The initial values for are given obviously by the known data (N(0)=25 for the St. Paul herd and N(0)=15 St. George herd). In order to choose a value for E(0) one has to consider the role that and indeed it has on the solution of the system 11). The maximum of the functions depends, both for its value and the placement, on E(0). If only non-renewable resources are available, then the function represents the value of the cumulative exergy available 0 is the total exergy available to the population at the to the population at time t, so initial time. A greater value of 0 corresponds to a smaller value of total exergy available, so that will be reached sooner: an increase of 0 will displace this maximum to the maximum of the left. Also, a change in the exergy available to the population will change its maximum 0 will cause a smaller value of this maximum. numerosity and a larger greater value of Summarizing we can say that greater values of 0 correspond to shifts on the left in the Cartesian plane of the maximum of and and to smaller values of these maximum. These is exactly what is observed comparing the numerosities of the two herds in the right of figure 4. In our case we have taken the values 3,2·10 -3 for the case of St. Paul island and 1·10 -2 for the St. George island. There could be different physical motives for this discrepancy, but the overall effect must have been such that a portion of the food reserves was not fully accessible in the case of St. George island.

5 - Conclusions In this paper we discuss the application of a population dynamics model previously developed in [12] to two different cases presenting very different behaviors in the dynamic of the populations as a response to the different type of environmental conditions in which populations live. We show that the case of a limit cycle dynamics of the population numerosity can be modeled by a time dependent external driving input to the system; the time dependence is due, in a real sense, to the action of the population on the environment. In such cases we demonstrate that the system is always attracted to the same curve independently from the initial conditions. This can be seen as a mathematical description of the notion of dynamical equilibrium developed in the realm of ecological systems (see for example [17] and references therein). Also, in section 3) we have shown that our model fits at first order in the expansion of the parameter the time series for the lynxhares interaction [10-11], having assumed the periodicity of this series. In the case of asymptotic stable equilibrium we have applied our model to the well documented dynamics of two reindeers populations in “closed” ecological niches. The non-trivial dependence of the population curves from the initial conditions (in particular 0 ), reflecting the ability of the populations to access to the food, is fully captured by the model and plays a significant role in the resulting dynamics. This is still a preliminary work needing to be refined and extended to other different species behavior, such as cooperation and parasitism, and even if the proposed model 190  

   

represents only a rough description of the high complexity of ecological systems, still it shows potential to be applied to very different situations.

Acknowledgments. One of the author (F.Z.) wish to acknowledge the financial support of the “Istituto Nazionale di Alta Matematica” (Italian National Institute for High Mathematics) as a Marie Curie scholarship holder.

References [1] E. Sciubba, F. Zullo: Is Sustainability a Thermodynamic concept?. International Journal of Exergy 2011 - 8, No.1 pp. 68 – 85. [2] E. Sciubba, F. Zullo: Exergy based population dynamics: a thermodynamic view of the sustainability concept. 2011. Journal of Industrial Ecology, 15, Issue 2, pp. 172–184, April 2011. [3] E. Sciubba: ‘Beyond thermoeconomics? The concept of extended exergy accounting and its application to the analysis and design of thermal systems’, Int. J. Exergy, Vol. 1, No. 1, pp.68– 84. [4] A.J. Lotka: Elements of Physical Biology, Baltimore,Williams & Wilkins Company, 1925. [5] B.E. Kendall, J. Prendergast, O.N. Bjørnstad: The macroecology of population dynamics: taxonomic and biogeographic patterns in population cycles; Ecology Letters, 1998, 1; pp. 160164. [6] M.P. Hassell, J.H. Lawton, R.M. May: Pattern of dynamical behavior in single-species populations. Journal of Animal Ecology; 1976; 45; pages 471-486. [7] N. C. Stenseth, R. A. Ims: Population dynamics of lemmings: temporal and spatial variation: an introduction. Pages 61-96 in N. C. Stenseth and R. A. Ims, editors. The biology of lemmings. Academic Press, London, UK. [8] E. Korpimäki, C.J. Krebs: Predation and population cycles of small mammals; Bioscience; Nov 1996; 46, 10. [9] K.R. Popper: The Logic of Scientific Discovery, Routledge, 14th Printing, 1977. First English Ed., Hutchinson, 1959. First published as Logik Der Forschung in Vienna: Springer, 1934. [10] C.J. Krebs, S. Boutin, R. Boonstra, A.R.E. Sinclair, J.N.M. Smith, M.R.T. Dale, K.Martin, R. Turkington: Impact of food and predation on the snowshoe hare cycle, Science. 1995; 269:1112–1115. [11] C.S. Elton: Voles, Mice and Lemmings. Oxford, U.K.: Clarendon; 1942. [12] V.B. Scheffer: The Rise and Fall of a Reindeer Herd, The Scientific Monthly, Vol. 73, No. 6 (Dec., 1951), pp. 356-362. [13] L. Hansson, H. Henttonen: Rodent dynamics as community processes, Trends Ecol. Evol. 3, pp. 195-200. [14] I. Hanski, L. Hansson, H. Henttonen: Specialist predators, generalist predators, and the microtine rodent cycle, Journal of Animal Ecology; 1991, 60, pp. 353-367. [15] J.D. Roth, J.D. Marshall, D.L. Murray, D.M. Nickerson, T.D. Steury: Geographical gradients in diet affect population dynamics of Canada lynx, Ecology, 88(11), 2007, pp. 2736-2743. 191  

   

[16] T. Klemola, M. Tanhuanpää, E. Korpimäki, K. Ruohomäki: Specialist and generalist natural enemies as an explanation for geographical gradients in population cycles of northern herbivores. Oikos 99, 2002, pages 83-94. [17] E. Mamontov: Dynamic-equilibrium solutions of ordinary differential equations and their role in applied problems, Applied Mathematics Letters 21, 320–325. [18] L.J. Palmer: The Alaska tundra and its use by Reindeer, U. S. Dept. of the Interior, Office of Indian Affairs, 55 pages (Mimeo). [19] G.D. Hanna: The Reindeer Herds of the Pribilof Islands, The Sci. Monthly, 15, Issue 2, pp. 181-186. [20] N. Tyler: Estimating the daily dry matter intake of Svalbard reindeer in late winter, Rangifer 7 (1): 29 – 32, 1987.

192  

PROCEEDINGS OF ECOS 2012 - THE 25 TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

The Influence of Primary Measures for Reducing NOx Emissions on Energy Steam Boiler Efficiency Goran Stupara, Dragan Tucakoviæa, Titoslav Živanoviæa, Miloš Banjaca, Srðan Beloševiæb, Vladimir Beljanskib, Ivan Tomanoviæb, Nenad Crnomarkoviæb, Miroslav Sijerèiæb a

Faculty of Mechanical Engineering, Kraljice Marije 16, 11120 Belgrade 35, Serbia, [email protected], CA b Vin a Institute of Nuclear Sciences, Mike Alasa 12-14, 11001 Belgrade, Serbia

Abstract: Within Electric Power Utility of Serbia 1991. a thermal power plant “Kostolac B”, power 2x350 MW started. Given the increased emissions of NO x in combustion products, it is in consideration to reduce it by introducing primary measures in terms of burner reconstruction, redistribution of secondary air by furnace height and introduction of cold flue gas recirculation. Applying these measures directly affects not only the combustion process in the furnace but also the work efficiency of steam boiler. Namely, if the flue gas temperature in furnace outlet is lowered below the corresponding values, it will cause the superheater and reheater not to achieve the designed steam parameters. On the other hand, the use of cold recirculating flue gases has a beneficial effect on reducing NOx and increases the work safety of the superheaters and reheaters, but reduces the steam boiler efficiency. In order to understand all these effects on the safety and efficiency of boiler operation, it was needed to create a software for thermal calculation of the observed steam boiler. Based on that software, this paper will display the impacts of introducing the measures for reduction of NOx emissions on a safe and efficient operation of the steam boiler as a whole.

Keywords: steam boiler, reducing NOx emissions, steam boiler efficiency

1. Introduction Modern society has big energy needs. Converting energy to a form useful to people has its consequences – harmful matter created in convertion process has adverse effects to humanity and environment. One of the harmful matter groups created in conversion process is nitrogen oxide group. Nitrogen oxides are inorganic chemical compounds where a link between nitrogen and oxygen is formed. The most frequent nitrogen oxides in the air are nitro-monoxide (NO) and nitro-dioxide (NO2) usualy common labeled as NOx. There are some other nitrogen oxides in the air, from which the most poluting one is nitrous-suboxide (N2O). Others, such as N2O3, N2 O4, N2O5, NO3, are not contained in the air in greater quanities. It is very important to say that most of the environmental protection regulations treat all nitrogen oxides as NO2. Although NO x emission level in our environment during lignite combustion is very low, but it still exceeds the European standard of 200 mg/Nm3 [1], which will become obligatory from 2016 in Serbia. Nitrogen oxides are very harmful compounds. Biggest source of nitrogen oxides with anthropogenic origin is fossil fuel, which is why we try hard to reduce nitrogen oxide emissions during fossil fuel combustion process. Traffic and energetics release 90% of total nitrogen oxide quantity relised by human activity. Nitrogen oxide adverse effect can be reduced easily by reducing total emission of nitrogen oxides. There are several different metods developed to reduse nitrogen oxide emission. Nitrogen oxides appear mostly through nitrogen oxidation from the air in combustion processes at high temperature ranges (''thermal NOx'') and nitrogen oxidation from the fuel (''fuel NOx''), which appears even at lower temperature ranges. In coal steam boilers, fuel NO x is dominant. Formation of 193

thermal NO x is directly dependent to local flame temperature, while formation of fuel NO x mostly depends on nitrogen content in fuel and reacting oxygen in the zone of particle burning. First group of methods consists of so-called primary methods. Primary methods are based on reducing nitrogen oxide emissions before and during their formation (before and during combustion process). Primary methods are much cheaper, but less effective. Most commonly used are recirculation of combustion products, usage of low nitrogen oxide concetracion burners (Low NOx burners), multistage combusting (OFA – Over Fire Air) or combination of those methods. Existing NO x reducing methods can be additionally improved and their efficiency increased by using numeric simulations. Numeric simulations compared to experimental researches are many times more cost-effective, easily doable and give satisfactory results. According to this, numeric simulations play an important role in designing new NO x emission reduction systems that will be used on facilities in development. In this way, even in early design stages, it is possible to predict the ammount of NO x emission, and whether the plant will satisfy the environmental regulations which are getting more and more strict. This paper cencerns the possibility of using primary methods for NOx reduction in outlet flue gases of energetic coal dust fired steam boilers, which refer to modification of the combusting processes in furnace of the energetic steam boilers in aspect of security and achieving work parameters of steam boiler. There was also a comparative analysis of diferent NO x reduction methods in aspect of steam boiler efficiency. Especially for these need, a numeric modeling of combustion process was made [2,3,4], along with a termal calculation code of the steam boiler in TE “Kostolac B” [10]. The selection of tested primary methods is made in a way that, in addition to their primary task (reducing NO x concentration in outlet gases), they do not violate the safe and efficient work of steam boiler facility.

2. Working conditions of the steam boiler There are two exploited steam boilers in the thermal power plant “Kostolac B” that burn lignite with lower heat value of 7326,9 kJ/kg. Performances of those steam boilers are: Main steam mass flow rate , D=277,8 kg/s Main steam pressure, ps=18,6 MPa Main steam temperature, ts=540 oC Reheated steam mass flow rate, D r=248,8 kg/s Reheated steam pressure, prs=4,375 MPa Reheated steam temperature, trs=540 oC Steam temperature at the reheated inlet, tr=334 oC Feed water temperaure, tnv=255 oC Simplified steam boiler disposition is shown in Figure 1. Flue gases made by coal combustion in furnace (1) stream over third superheater stage (3), second reheater stage (4), second superheater stage (5), first reheater stage (6) and economizer (7), and than turn into sheet duct (8), at which outlet there are two air preheaters. Flue gases are then released into the atmosphere. Steam boilers in TE “Kostolac B” (nominal power 2x350 MW) are tower shaped with forced circulation. Furnace dimensions are 15,1 x 15,1 x 43,0 m, with solid state dross drainage and rost. Boiler is stoked with coal dust, using eight tangentially placed burners, each connected to its own fan mill (9). Burners are separated into four levels by height: lower and uper main burner (10) and two burners for coal laden vapour above (11) for burning smaller fractions of coal dust. In case of maximal permanent block power (100 % of the load) with 7 mills working, raw coal consumption is 194

119,13 kg/s, and hot air flow is 1050 103 Nm3/h. Temperature of mill gases is 200 OC, coal dust flow per burner is 10,384 kg/s and transport fluid flow per burner is 43,726 kg/s. Hot air temperature is 288 OC, secondary air flow per burner is 38,2 kg/s, tercial air flow trough rost is 15,18 kg/s. Guaranteed fuel is lignite Kostolac - Drmno, technical analysis: moisture 43,93 %, mineral mater 22,25 %, volatiles 21,39 %, fixed carbon 12,43 %; elementary analysis: carbon 22,46 %, hydrogen 2,12 %, oxygen 7,7 %, nitrogen 0,9 %, burnable sulfur 0,64. Moisture content in coal dust is 8,83 %. Coal dust particles density is 1300 kg/m3. Operational requirements data in project regime, working fuel and coal dust caracteristics are shown in 2 . Based on sifter analysis, RozinRamler distribution and numerical experiments, the diameter of mono dispersal coal dust particle was taken as d p=150 m, for simulation purposes. During the study, a developed model of formation and destruction of NO x was beeing used, verified by comparison to available measurments of NO x emission in thermo energetic facilities of TE “Kostolac B” and incorporated with earlier developed complex model of the processes in furnaces [3] of subjected blocks. For the analysis of impact the NO x reduction measures have on the efficiency of the entire boiler plant, and the efficiency and safe work of the superheater, for achieving the designed parameters of steam, a thermal calculation code was used [10], based on Norman’s method [11].

Figure 1. Disposition of the steam boiler in TP Kostolac B: 1. Furnace; 2. First superheater stage; 3. Third (output) superheater stage; 4. Second (output) reheater stage ; 5. Second superheater stage; 6. First reheater stage; 7 Economizer; 8. Sheet duct; 9. Fan mill; 10. Main burners; 11. Burner for coal laden vapour.

3. Numerical study of the possibility of NOX reduction 195

Using the developed model of formation and destruction of NOx [2,3,4], the impact of primary methods, namely, differently organized combusting processes in furnace, on reduction of NO x emissions is numerically tested. Using a numerical simulation, the influence of different distribution of mill gases and heated air by burner levels, the impact of OFA vents and an effect of outlet gas recirculation from the steam boiler outlet on nitrogen oxide emission, was tested.

3.1. Research of the effects of diferent mill gases and air distribution to burner levels on reducing NOX Influence of the distribution of mill gases and air mixture on reducing the nitrogen oxide concetration was observed for three test cases (TS 1-3). In all three test cases steam boiler worked with 6 mills and designed parametars with nonworking mills on opposite walls so that the temperature field and concentration were very close to simetrical. Varied parameters are shown in Table 1. TS-1 is a project operating mode of the boiler, therefore it is taken as reference in studying the impact of the primary methods on NO x emission. Numerical results for gas mixtures temperature field, and O2, HCN and NO x concetration field in the furnace of the TE “Kostolac B” steam boiler for project conditions (TS-1) are presented in Figures 2. and 3.

Figure 2. Gas thermal field and O2 concetration field in furnace TEKO B for TS-1;

Figure 3. Concetration fields of HCN i NOx in furnace TEKO B for TS-1; 196

Numerical results show the dependence of obtained NOx concentration fields on gas temperatures, flame position, and field of reactants (HCN in this case) and oxygen concentration in furnance space, which, by the homogeneous reactions, produce burnable NO x. There are also reactions of NOx reduction with HCN. Thermal NO x appears in a narrow area of maximum local temperatures in the furnace (T=1650K-1800K), which was expected. Thermal NO x concetration in total NO x emission is only a few percent. Since the fuel NOx is much more represented then thermal NO x (for the temperature range in the considered furnace), dependence of fuel NOx predominantly determines the caracter of total NOx concetration field. Figure 3 suggests a significant effect of HCN concentration (as volatiles intermediar), and therefore the nitrogen content in coal from which the HCN comes from, to NOx content. One, narrow zone of high NOx concentration corresponds to a maximum content of HCN is the place of introduction of most of the fuel through the lower main burners. The second, broader zone of high NOx concentrations is noted along the flow (upwards) and corresponds to areas of intense chemical reactions of HCN consumption and forming of NO x. In contrast to thermal NO x the content of fuel NO x (therefore the total NOx) is less affected by temperature, but (besides the content of nitrogen in the fuel) it is greatly influenced by the relation of air and fuel, more precisely, the excess air (oxygen concentration field). That clearly follows from the comparison of Figure 2. with Figure 3. The NOx concentration field does not only follow the temperature field (and HCN field) but, even more, the O2 field. Despite the high temperatures, furnace central area does not have high concentration of NO x because the oxygen is depleted by intensive, and relatively quick reactions of fuel combustion [5]. Numerical results for NO x emission in the subject furnace were satisfactory matched with the results of periodic emission measurements in the TE "Kostolac B" blocks. Emission values specified by the model refer to average values at the furnace outlet, converted so they can be compared with the measurements of the steam boiler (and values defined in the standard). Numerical result of the NO x concentration at the end of the furnace is 414.9 mg/Nm3. Operating mode TS-2 represents a possible case of redistribution of mill gases and combustion air, with all other input parameters unchanged, which is more preferable from the aspect of NO x concentration. Input parameters of TS-2 mode are shown in Table 1. Figure 4. shows the appearance of gas mixture temperature field and nitrogen oxides concentration per longitudinal section of the furnace. Numerical result of the NO x concentration at the end of the furnace is 343.6 mg/Nm3 which is a significant improvement of the nitrogen oxides concentration in relation to the TS-1 mode.

Figure 4. Gas temperature field and NOx concentration field in furnace of TEKO B for TS-2; Values of nitrogen oxides' concentration for three examined cases of the fuel and air distribution per burner levels are given in Table 1. 197

Operating mode TS-3 represents a reorganization of additional fuel intake per burner level with large amounts of fuel through the main burners. This operating mode would be more advantageous over the desired criteria to reduce the concentration of NO x. Input parameters of TS-3 mode are shown in Table 1. Figure 5. shows the appearance of temperature field of gas mixtures and concentrations of nitrogen oxides per longitudinal section of the furnance. The numerical result for the concentration of NOx at the end of the furnace is now further enhanced, and is 337.6 mg/Nm3 which represents this operating mode interesting for further analysis from the viewpoint of maintaining boiler operating parameters and cost-effectiveness of this solution. In this way the numerical experiment has shown that by the favorable mill gases and distribution of combustion air can affect the reduction of nitrogen oxide concentration in the gas mixture. Table 1. The distribution of mass fuel flow and combustion air per burner level; Fuel distribution per burner level Transport Secondary [%] fluid air through Test tizl through Coal laden main case Main burners [oC] main vapour burner burners (TS) burners [%] Lower Upper Lower Upper [%] 1 45,5 24,5 19,5 10,5 57,0 67,8 1021 2 35,2 44,8 8,0 12,0 60,3 67,8 1033 3 43,2 43,2 7,9 5,7 57,4 67,8 983

NOx Emission [mg/Nm3] 414,9 343,6 337,6

Figure 5. Gas temperature field and NOx concentration field in furnace of TEKO B for TS-3;

3.2. Examine the influence of OFA vents on NOX reduction As one of the most important primary measures for reducing nitrogen oxide concentration in the mixture of gases on dust coal-fired boilers, the multilevel air intake (air stagging) is being used. This kind of measure is actually a two stage injection or two stage combustion. With this system, the total amount of the secondary combustion air is divided in two parts so that approximately 70 90% of air is injected through the burners and thereby the lower flame temperature and richer fuel mixtures are achieved in this zone. These two conditions allow that level of nitrogen oxides emission in burner level is less than with the classical system. The remaining 10 - 30% of the combustion air is blown in through special air vents, which are located above the burner (OFA), in 198

order to achieve complete combustion [8]. In this way, in the level of OFA vents, a poor mixture zone is achieved, where the emission of nitrogen oxides is less than in conventional systems. The idea is actually to delay the combustion of certan amount of fuel until the area of OFA vents where, due to the local increase of excess air, a lower temperature is achieved, which results in a lower NOx emission [5]. This system is simulated numerically for discussed test cases. Only cases with the most optimal OFA vent parameters for the considered operating mode are shown. Figure 6. shows temperature fields of gas mixture and NOx concentrations for TS-1 with two-stage air injection. OFA vent has the same width as the burner vent, height of 1 m, and 3 m away from the burner for coal laden vapour. The optimal flow in this case is 10% of air. In this way there was an additional reduction of nitrogen oxide emissions, to 393.8 mg/Nm3, or 5.1% less compared to the set distribution of the working regime TS-1.

Figure 6. Gas temperature field and NOx concetration field in furance of TEKO B for TS-1 with OFA vent;

Figure 7. Gas temperature field and NOx concetration field in furance of TEKO B for TS-2 with OFA vent;

199

Figure 8. Gas temperature field and NOx concetration field in furance of TEKO B for TS-3 with OFA vent; Test case 2 is, in terms of concentration of nitrogen oxides at the furnace outlet, improved by introducing OFA vents on 6 meters from the burner for coal laden vapour. Vents are the same size as in the previous case, but the air flow through them is increased to 20%. This resulted in reduction of NO x concentration by 11.8% from the original TS-2 and now stands at 303.2 mg/Nm3. The temperature fields and NOx concentration are shown in Figure 7. In test case 3, the identical vents are installed, as with improved TS-2, but with an air flow rate of only 5%. In this case, the concentration of NO x is reduced by 6.3% compared to the original TS-3, and is now 316.5 mg/Nm3 in the output section of the furnace. New balance is shown in Figure 8.

3.3 Examine the influence of recirculation from the end of the boiler on reduction of NOx NOx emission reduction can be achieved by recirculation of cold flue gases from boiler exits back to the combustion process. Flue gases are brought back into the primary combustion zone and in this way emission of nitrogen oxides is beeing reduced, by two mechanisms [6]: Flue gases acts as an inert component in the fuel-air mixture. Additional mass of cold flue gas is heated in the flame, causing a reduction of the flame temperature, while reducing the created amount of thermal nitrogen oxides, Introduced flue gases reduce the oxygen content in the primary combustion zone, thus reducing the created amount of nitrogen oxide. Main disadvantage of reducing the flame temperature at flue gas recirculation is a lower overall combustion efficiency. Besides that, there is the problem of flame stability, emissions of CO and solid matter [7]. The injection of flue gas can be achieved by mixing with combustion air and fuel prior to entering the combustion zone, or may be directly introduced into the flame zone. This system enables, from technical side, an easily applicable primary measure of nitrogen oxide reduction. Boilers of the TE "Kostolac B" operate on individual coal dust preparation system with direct blowing and drying at close process where the certan recirculation from the end of the boiler is predicted, in purpose of inertisation of mill gases and reduction of the risk of explosion [9]. This system enables, from technical side, an easily applicable primary measure of reduction of nitrogen oxides. The cold flue gases from the end of the boiler are brought into recirculation head where they mix with primary air, fuel and flue gases from the end of furnace before they are introduced into the fan mill. 200

The possibility of reduction of the flue gas recirculation in relation to the projected, in terms of favorable NO x concentration, and the safe and efficient operation of the plant as a unit is considered in the test cases. This primary measure was tested by a numerical simulation with 0%, 4% and 8% of recirculated gases from the end of the boiler at on one of the working conditions measured in 2011. All prior cases were conducted with the project gas recirculation from the boiler end, which was 4.9% of the total output of flue gases (nominal load 1000 t/h). Figure 9. shows the temperature and concentration fields of HCN, O2 and NO x for the three test cases of recirculation of flue gases of 0%, 4% and 8%. In the first case, when there is no recirculation, obtained content of NOx was 748.0 mg/Nm3 , in second case, when the simulated recirculation was 4%, it was was noted that emission was reduced by 14.1%, to 642.7 mg/Nm3, while in the third case the recorded reduction was 23.9% compared to the reference case without circulation, and was 569.1 mg/Nm3. In Figure 8 the effects of the characteristic mechanisms for this type of primary measures are noticeable, which are more expressed with increasing of recirculation flow.

4. The influence of the examined primary measures on operating effectiveness of the boiler As criteria for selection of optimal primary measures (combustion modifications) to reduce NO x emissions from the standpoint of the need for efficient operation of the complete boiler plant, as well as efficient and safe operation of super heater in terms of achieving the designed parameters of steam, we can single out the following: a steam boiler efficiency degree, an efficient and safe operation of the super heater and reheater, in terms of achieving the designed parameters of steam and minimum required amount of water injected into the lines of main steam and reheater steam. In Table 2 are given the average temperatures of flue gas on the combustion chamber exit, obtained by a mathematical model for three selected test cases with low NO x emission with and without OFA vent. Based on these temperatures, related to the nominal strain of the boiler, adjusted to the thermal calculation of the furnace in order to get identical values upon which implemented the heat boiler calculation for the projected fuel or warranted coal for the boiler. The values that need to be analyzed according to the temperature changes of flue gases at the end of the furnace are shown in Table 2. Table 2. Thermal calculation results according to the temperature of flue gasses on the boilers exit;

Title

Label

Unit

Test case with diferent fuel and air distribution by stages

Test case with OFA vent

TS-1

TS-2

TS-3

TS-1

TS-2

TS-3

1021

1033

982

996

1000

974

0

3,078

4,253

0

Flue gass temperature in the furnance exit

t l”

[oC]

Injection of the water in main steam line

DHs

[kg/s]

ts

[oC]

540

540

536

540

540

525

DHr

[kg/s]

4,210

5,487

0,783

1,882

2,182

0,532

Main steam temperature Injection of the water in reheated main steam line

201

15,505 18,822

Main reheated steam temperature Fuel consumption Temperature of the flue gass on boilers exit Heat loss by the flue gass on boilers exit Boiler efficiency

trs B

[oC] [kg/s]

t iz

[oC]

158,9

159,3

157,7

158,2

158,4

157,1

q2

[%] [%]

11,66 85,06

11,69 85,02

11,56 85,15

11,60 85,12

11,65 85,09

11,42 85,29

k

540 540 540 540 540 534 120,35 120,88 118,46 119,45 119,5 118,12

With the high temperatures of flue gases at the end of the combustion chamber, the third stage main steam superheater PP3 (Figure 1. - by the output of steam flow), receives an increased amount of heat due to the rise in the average logarithmic temperature difference of heat transmitter and receiver. In Table 2 is noted that the flue gas temperature of 1033 °C injecting in main steam flow line is 18.822 kg / s of water to maintain the exit temperature of main steam flow of 540 °C. By reducing the temperature of the flue gases at the end of the combustion chamber, main steam superheater PP3 exchange a small amount of heat which leads to a reduction in the required flow of water for injecting in. The problem occurs at temperatures of flue gases at the end of the furnace below 990 °C, since it can not achieve the design parameters of the main steam of 540 °C, which are the primary measures applied to the TS-3, which gave good results in terms of reduced concentrations of nitrogen oxide, are disqualified as a possible solution. Above the third degree of main steam superheater, by the gas flow, the second stage of main steam reheater is set NP2 (output by the steam flow) at high temperatures of flue gases also receive an increased amount of heat. This percentage increase is slightly less than the increase that occurs in the third degree of main steam superheater. Table 2 shows that the lowering the temperature of flue gases at the end of combustion chamber, water injection in the reheated steam line are reduced, but that in all the considered cases, however, achieved subsequently superheated steam temperature at the outlet of 540 °C. For this boiler regulation of reheated steam temperature can only be achieved by injecting water between the two levels (there is not any bifluks or trifluks). By increasing amounts of water for injecting into the reheated steam line (DHr) efficiency of the block is reduced (because the amount of injected highpressure bypasses the turbine), but also increases fuel consumption (due to the increased amount of heat needed for local heating of injected feedwater DHr), which can be seen displayed on Table 2. In test case no. 2 there is a need for higher amount of water for injecting in into the line reheated steam, which in this case is an acceptable price for the sake of a significant reduction of NO x emissions. By the increased heat exchange in the third stage of the steam superheater and the second stage of steam reheater at elevated temperaure of flue gases exiting the combustion chamber, flue gas temperature in front of and behind other heating surface temperatures are approaching the designed. Table 2 clearly observed that the temperature of flue gases at the end of the boiler, and therefore the boiler efficiency, slightly change the temperature of flue gases at the end of the combustion chamber, by enabling those measures applied in TS-2 are acceptable from the standpoint of efficient and safe operation of boiler the plant. This test case can be further improved by introducing OFA, so that the new reorganized combusting besides favorable concetration, further reduces the temperature of gas at the end of the combustion chamber, reduces the required amount of injection and increases efficiency. So the test case 2 with OFA vent is considered the most optimal solution. Influence of recirculation in terms of boiler plant efficiency is shown in Table 3. Giving the thermal budget for running the test cases without recirculation and with recirculation of 4% and 8%. Increasing the recirculation of flue gases from the end of the boiler and further cool the furnace and lowers the temperature of gases exiting the combustion chamber. Also, by increasing the recirculation of cold flue gas NOx emission is reduced, however, significantly increases the temperature of flue gas exiting the boiler, leading to a reduction in boiler efficiency. 202

Flue gas temperature at the end of the combustion chamber is lowered, but due to the increased flue gass flow through the boiler (due to cold gas recirculation) the amount of water that is injected to regulate main and reheated steam temperature increases, which again leads to a reduction steam block efficiency and increased consumption of fuel. Table 3. Steam boiler termal calculation results for diferent gas recirculation from at end of the boiler;

Test case

0% recirculation

4% recirculation

8% recirculation

Flue gass Main and temperature Water Water Temperature reheated Fuel injecting in of the gass that Output gass at the end of injecting in consumpt steam the main the reheated exits the heat loss, the temperature ion, steam line, boiler, [%] combusting steam line, [kg/s] , [kg/s] [kg/s] [ oC] chamber, o [ C] [ oC]

Boiler efficency, [%]

Emission reduction NOx comparing to the emission case without gass [mg/Nm3] recirculation (%)

1087

28,803

540

9,719

103,39

160

10,36

86,56

748,0

-

1057

24,485

540

9,109

103,71

166

10,78

86,14

642,7

14,1

1028

19,753

540

8,278

103,93

172

11,18

85,74

569,1

23,9

203

recirculation of the boiler exit recirculation of the boiler exit recirculation of the boiler exit flue gases: 0% flue gases: 4% flue gases: 8% Figure 9. Temperature fields and concetration fields of HCN, O 2 i NO x for 3 recirculation cases;

5. Conclusion Based on the numerical simulations performed to investigate the possibility of reducing NO x emissions, it was observed that the distribution of coal dust per burner stages and secondary air significantly affects the emission of NO x and the exit temperature of flue gases at the end of the furnace. More coal dust passing through the main burners (85%), and control of local excess air in the burner zone, by redirecting up to 20% of heated air from the main burners to OFA vents, have a significant impact on the reduction of NO x emissions. Cold flue gas recirculation will reduce the concentration of nitrogen oxides, but it will always result in a lower efficiency (Table 3). Numerical simulations show that optimizing the combustion process can significantly reduce NOx emissions, and keep the temperature in furnace in the required range. In order to achieve optimum combustion process it is necessary to determine the proper distribution of coal dust and heated air to individual burners and burner stages. This can be done without large investment costs. In this way it is possible to achieve reduction of emissions by 20 - 30% without structural changes of the boiler, only by combining the tested methods of combustion process optimization. Based on the analysis performed in the work, it can be concluded that TS-2 with OFA vents gives the best result from both considered aspects, with the possibility of increasing the recirculation from designed 4.9% to 8% which would further reduce the concentration of nitrogen oxides in flue gases at the end of combustion chamber. Based on the thermal calculation of the boiler, it can be concluded that the optimum temperature of flue gases at the end of combustion chamber should be in range of 990-1010 °C, in order to provide for safe operation of third steam superheater stage PP3, 204

namely to provide the necessary main steam temperature of 540 °C. At this temperature range, water injection in the reheated steam stream is minimal. Test results show that the European standard can not be achieved using primary measures only, but with proper selection of those measures, it is possible to significally reduce the starting concentration of NO x for use of secundary measures which require additional investment and exploitation expenses.

Acknowledgments This work has been supported by the Republic of Serbia Ministry of Education and Science (project: “Increase in energy and ecology efficiency of processes in pulverized coal-fired furnace and optimization of utility steam boiler air preheater by using in-house developed software tools”, No. TR-33018).

References [1] Directive 2010/75/EU European Union - limit of emissions of harmful substances into the air from large furence. [2] Belosevic S, Sijercic M, Tucakovic D, Crnomarkovic N. A Numerical Study of a Utility Boiler Tangentially-fired Furnace under Different Operating Conditions, Fuel 87, 15-16, pp. 33313338. [3] Belosevic S, Sijercic M, Oka S, Tucakovic D. Three-dimensional modelling of utility boiler pulverized coal tangentially fired furnace, INTERNATIONAL JOURNAL OF HEAT AND MASS TRANSFER 49, 19-20, p.p. 3371-3378. [4] Belosevic S, Sijercic M, Crnomarkovic N, Tucakovic D. Numerical prediction of pulverized coal flame in utility boiler furnaces, Energy & Fuels 23, p.p. 5401-5412. [5] Belosevic S, Sijercic M, Crnomarkovic N, Zivanovic T, Tucakovic D. Possibility of introducing the primary measures for reducing nitrogen oxide emissions from power boilers on pulverized lignite. Belgrade, Serbia: Laboratory for Thermal Engineering and Energy, Institute of Nuclear Sciences Vinca, and the Center for Thermal Engineering - Department of boilers, Faculty of Mechanical Engineering, University of Belgrade, 2011. [6] J. Dukovic, V. Bojanic. Air pollution - a term, condition, resources, control and technological solutions. Institute for Protection and ecology - Banja Luka, Bosnia and Herzegovina, 2000. [7] J. Baltasar, M. Carvalho, P. Coelho and M. Costa, Flue gas recirculation in a gas-fired laboratory furnace: measurements and modelling, Fuel 76, pp. 919-929 [8] Tobin, D. Moyeda, W. Zhou, R. Payne, Application of Layered Control Technologies to Significantly Reduce NOx Emissions from Coal-Fired Boilers, GE Energy, 2ndU.S.-China NOx Workshop Dalian, China, 2005. [9] Brkic Lj, Zivanovi T, Tucakovic D. Steam boilers, Faculty of Mechanical Engineering, University of Belgrade, Belgrade, Serbia 2010. [10] Tucakovic D, Zivanovic T, Stevanovic V, Belosevic S, Galic R. A computer code for the prediction of mill gases and hot air distribution between burners sections at the utility boiler, Applied Thermal Engineering 28, 17-18, p.p. 2178-2186. [11] Brkic Lj, Zivanovic T, Tucakovic D. Thermal calculation of steam boilers Belgrade, Serbia: Faculty of Mechanical Engineering, University of Belgrade, 2010.

205

PROCEEDINGS OF ECOS 2012 - THE 25 TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

The LETHE CITY CAR of the University of Roma 1: Final proposed configuration Roberto Capataa, Enrico Sciubbab a b

University of Roma Sapienza , Dept. of Mechanical and Aerospace Engineering, Italy, [email protected] University of Roma Sapienza , Dept. of Mechanical and Aerospace Engineering, Italy, [email protected]

Abstract A longstanding interest of the Authors’ research group at UDR1 was the design, development and fielding of a road prototype of a new concept of Hybrid Series vehicle, endowed with a small Gas Turbine set as a thermal engine. This solution offers several advantages with respect to traditional internal combustion engines and even to the existing generation of Hybrid propulsive systems,: a reduced engine weight and size, lower emissions, substantially extended range, ease of maintenance, and more efficient braking energy recovery. In the LETHE (Low Emissions Turbo-Hybrid Engine) the GT does not directly provide traction, but serves solely as a battery pack recharger. The vehicle is, in all respects, equivalent to a purely electric vehicle, except for the presence of an on-board recharger. Much care was placed in the design phase in the quest for an “optimal” design: first of all, an original method for identifying the most convenient degree of hybridization (ratio of the installed power of the battery pack and that of the GT) was defined and formalized, so that the resulting power balance between the two units satisfies the main design specifications, namely that of guaranteeing a practically acceptable operational life of the battery package while enabling the vehicle to complete a typical city mission (about 25-50 km) in a purely electric mode and without recharge. This paper presents a review of the previous conceptual and design results and describes in detail a possible road prototype configuration (weights, packaging of the units within the body of the vehicle, logic control unit, GT- and electric motor size and power, battery package characteristics). Some discussion is also devoted to the foreseeable impact of the deployment of a LETHE fleet on the mid-range scenario of the Italian urban transportation system.

Keywords: Hybrid Vehicle, Ultra-Micro Gas Turbine, Vehicle emissions, CO 2 abatement, Transportation Economics

1. Introduction: a brief review of existing hybrid vehicle concepts and of current market opportunities In the last decade, governmental incentives and the ever stricter emissions regulations have prompted some of the largest world automakers to allocate resources to the study, design, development and production of hybrid vehicles, which offer undisputed advantages in terms of emissions and fuel consumption with respect to traditional internal combustion engines. In fact, true hybrid engines are substantially smaller than conventional ICE, because they are designed to cover the vehicle’s “average” power demand, which ensures proper traction for about 99% [15] of the actual driving time, and is exceeded only for prolonged mountain drives and instantaneous accelerations. When excess power is needed above this average, the hybrid vehicle relies on the energy stored in its battery pack. Hybrid cars are often equipped with braking energy recovery systems that collect the kinetic energy lost in braking, which would be dissipated into heat otherwise, and use it to recharge the battery. Smaller sizes and an (almost) constant operational curve lead to lower emissions. Moreover, a hybrid vehicle can shut down completely its gasoline engine and run off its electric motor and battery only, at least for a limited operational range: this “mixed operation” increases the net mileage and releases a substantially lower amount of pollutants over the vehicle lifetime. Due to market demand though, current commercial hybrid vehicles (HV) are mostly passenger hybrid cars equipped with a traditional ICE and an electric motor coupled in parallel. The thermal engine is normally oversized with respect to the average power, and the 206

surplus power needed during rapid acceleration phases is supplied by the electric motor: as a consequence, fuel savings are limited, as are global emissions, and the electric range is severely limited.

2. The LETHE@ concept The series hybrid configuration developed by the authors’ research group [2, 3], nicknamed LETHE@, is a vehicle in which two natural gas fuelled small turbogas sets are coupled to high speed electrical generators and a lead-acid battery package: the vehicle can operate in electric-only mode if requested, or in hybrid mode, where the gas turbine and the battery package operate together to satisfy the power demand [3,4]. The traction is fully electric in either operational mode. In the hybrid vehicle scheme discussed in this paper, the electronic vehicle management unit (“VMU”) controls ignition and on-off switching under a Load Following logic. The VMU decides at each instant time how much of the energy produced by the GT reaches the battery package or the electric engine directly. In addition, the electric motors can also act as brakes, recovering much of the energy that is otherwise lost. In order to maximize the recovered energy and to avoid possible battery overloading, an additional dynamic storage unit has been included: a relatively small flywheel capable of storing the excess power from the regenerative braking and of releasing it at a later time according to the instantaneous power demands. The VMU performs its energymanagement task on the basis of a certain number of instantaneous mission parameters: the batteries may thus provide or absorb the difference between the energy requirements of the vehicle and the GT energy production. The generator acts as a starter for the GT as well. A continuous GT control can be enforced via fuel flow control and/or employing a variable geometry GT. Since GT power modulation is affected by a substantial efficiency penalty at off-design conditions, the fuel flow control is coupled with a variable-stator turbine and the inlet guided vanes (IGV) blades for the compressor. As any other system, the GTHV has advantages and drawbacks. The following parameters ought to be considered when selecting/designing such a system: It is of compact size and can be comfortably mounted in the engine compartment of a sedan; Both the micro turbines and the electric engine have a very high power-to-weight ratio; The GTHV attains a very high fuel economy; The GTHV has a lower emission level, with effective multi-fuel capability; There is the possibility of improving the overall vehicle design due to weight and size savings; All components have a high reliability; The battery package has a rather low power-to-weight ratio; The state of charge (SOC) trend during any mission must be monitored to avoid overcharge and excessive discharge of the battery pack; The GT may be subjected to several ignitions during a mission, which negatively affects its mean-time-between-failure (MTBF); There is the necessity of monitoring and satisfying the instantaneous vehicle total power demand.

3. The degree of hybridization The mechanical power in an series HV vehicle is typically supplied by one electric motor (EM), so that, from the traction point of view, the vehicle is in fact an electric one. The choice of the EM is a direct function of the required performance. Once the maximum required traction power is fixed, then the total power source supplied by the ICE and battery package can be calculated from the overall mission energy balance. Thus, the Hybridization Degree HD can be calculated as the ratio between the GT power and the total installed power (GT and battery package). 207

HD = PGT / PGT+BP (1) Our design target is to attain the minimum possible HD that still guarantees a good driveability under all possible conditions.

4. Simulation Several numerical tests have been carried out to compute the vehicle performance, in two different driving missions: a combination of 10 consecutive urban cycles ECE15 and a “complex driving mission” composed of 4 consecutive extra-urban cycles EUDC and 72 minutes of continuous highway drive at 120 km/hr. Each mission has been simulated for each of the two concept cars studied here: a “city-car” and a standard passenger sedan. The simulation computes the power balance on the basis of the imposed wheel speed and vehicle characteristics [1,2], and determines the power supplied by each system component. This process is repeated with a 1s interval, assuming that within every time interval, the power, the speed, and all other significant parameters remain constant. As mentioned above, the GT set is switched on when the SOC is lower than a set point (0.6), and switches to idling or partial load mode when the SOC reaches the maximum set point (0.8). In real operation, a manual override must also be provided, but this was not considered in the calculations. The GT load management protocol is based on the assumption that the GT sets can operate, without substantial efficiency loss, between 70% and 110% of their nominal power. Each simulation, consists of assigning first the number of modules in the battery package, then the installed power, and finally the GT power: these three values must satisfy the limitations imposed by the above- mentioned criteria of maximum power demand and maximum absorbable battery power [7]. The GTs nameplate power was iteratively adjusted until the minimal fuel consumption was obtained. This heuristic procedure was also iterated by increasing the number of battery modules, with a consequent correction of the total vehicle weight. The vehicle design specifications (Table 1) are the same as those adopted in previous papers [2,3,4]: Table 1. GT Hybrid Vehicle (GTHV) Design Specifications Wheel rolling radius R = 0.265 m Vehicle width b = 1.7 m Vehicle height H = 1.4 m Net front area Sf = 2.142 m2 Area ratio (Sf/Stot) = 0.9 Aerodynamic drag coefficient cx = 0.25 Tire rolling friction coefficient fr = 0.015 Vehicle mass m = 1200 kg Equivalent mass me = 1240 kg Air density = 1.18 kg/m3 Air intake temperature T = 300 K Minimum SOC 0.6 Maximum SOC 0.8

EEC Directive 90/C81/01: this is a series of Regulations that prescribe both the emissions limits (adjusted every year) and the methods for testing and qualifying passenger and commercial vehicles. The test driving are in one urban cycle (European Cycle Emission) and an extra urban driving mission (Extra Urban Driving Cycle)

208

5. Results of the simulations Eight different computer simulations have been performed (2 types of mission respectively simulated with 2 types of logic, and 2 types of battery recharge limit BRL). Urban

Complex mission

cycle

Logic A

Logic B

BRL BRL

BRL

2C

C

C

Logic B

Logic A

BRL

BRL

2C

C

BRL

BRL BRL 2C

2C

C

Fig.1 . Scheme of the performed simulations The choice of the optimal configurations, within those several simulations, is a heuristic balance between the relative advantages and drawbacks of the following parameters: Total gross weight of the battery package; SOC trend during the mission; Number of GT ignitions during the mission; Instantaneous coverage of the total demand power of the vehicle; Size of the several devices (GT, battery package, flywheel)

6. Vehicle Hybridisation In the vehicle hybridisation process, once the initial calculations have been completed, as indicated in previous works [1, 2], each component of the Lethe® vehicle is then individually designed. For a 30 kW electric motor, the overall dimensions of the main components are reported in table 2.

h L

l

L

h

D

L

l

Fig. 2. Asynchronous motor da 30 kW; L = 315 mm, D = 264 mm

Fig. 3. Inverter; L = 410 mm, l = 340 mm, h = 138 mm

Fig. 4. GENESIS ® battery module; L = 200 mm, l = 170 mm, h = 170 mm

h L

D

L

l

Fig. 5. Gas Turbo-generator; Fig. 6. Fuel tank; L = 880mm, L = 465 mm, D = 270 mm D = 200 mm 209

L D

Fig. 7. Regenerator; L = 340 mm, l = 215 mm, h = 120 mm

Table 2. Components dimensions and weight Component Dimensions [mm] Weight [kg] Electric motor Ø264 x 315 80 Inverter 410 x 340 x 138 15 Battery module 200 x 170 x 170 15 GT device Ø200 x 465 25 Fuel tank Ø270 x 880 37 Regenerator 340 x 215 x 120 12 Ø200 x 270 16 (Fly wheel) Total weight 200 +(16)

6.1. Note about the selection of the Gas Turbine The design of the GT units were not a subject of this study, therefore the sizes and weights of the unit were not expressly calculated. However, in this study, is the energy balances were performed on the basis of the known characteristics of the 30 kW Capstone turbo-generator C30HEV [14], therefore the packaging reflects an excessive size, because the actual optimal degree of hybridization is about xx and would indicate the need for a xx kW turbogas set. Indeed, we assumed that the temperature (1300 K) and speed (100000 rpm) were the same as the C30HEV, therefore as an initial approximation, the sizes should be scaled respectively by a factor of 1/3. With the shape and size of the C30HEV [14] established, such an “ideal” GT group was repackaged as shown in figure 4. Its weight was also evaluated in excess, approximating it to that of a single steel cylinder ( steel = 7.87 kg/dm3).

7. The LETHE@ City Car The weights and dimensions of all components are summarised in Table 3. The battery weighs 90 kg (6 modules) and is placed underneath the rear seats. The gas tank is placed in the aft section while all remaining components are housed in the front section (Figure 8). This configuration allows a weight distribution of 146 kg on the fore-axle and 127 kg on the rear axle, for a 53/47 ratio (Figure 9).

Table 3. City Car Configuration W [kg] Dimensions [mm]

Component

Batt. Pack (6 mod.) E. Motor/Generator Inverter GPL Fuel Tank Fly wheel Fly wheel motor GT 30 kW Regenerator

90 80 15 37 3 13 23 12

510x400x185 Ø264x315 410x340x138 Ø270x880 Ø200x270 Ø200x465 340x215x120

The flywheel’s task is that of smoothing sharp braking, downward slopes etc.. It was sized so as to obtain a compromise between storable energy and volume. Once the amount of energy that the flywheel had to absorb (20 kJ) and the maximum rotation speed (20000 rpm) were set, the weight and disk radius were computed using standard formulae [4].

210

Fig. 7. Views and main dimensions (in mm) of the LETHE® City Car configuration

146 kg

127 kg

Fig. 8. Weight distribution in the LETHE® City Car configuration

8. Calculation of emissions for the LETHE @ vehicle Pollutants emissions are linked to several factors (such as the type of fuel, the combustion

temperature, the air/fuel ratio), and vary depending on operational conditions. The small amount of emissions data available on emissions at off-design conditions refer to large TG plants, where the combustion characteristics are different (pre-mixer, higher compression ratio), while the available data for small plants are only available for stationary conditions. As an initial approximation, emissions from the LETHE@ under the tested missions were estimated by comparison with emissions from the Capstone turbo-generator C30HEV [15], having presumed that our GT unit has the same inlet temperature in the turbine, same compression ratio and same speed. The Capstone data are expressed in g/kWh and refer to a fuel mass flow rate of 2.36 g/s. Table 4. Emissions from the Capstone C30HEV unit Emissions [g/kWh] NOx

CNG 0,194

Propane 0,396

HC

0,313

0,313

CO

0,306

0,134

PM

0,003

0,003

It was also assumed that the GT unit is regulated by variation of the fuel capacity, maintaining the turbine inlet temperature constant. For this reason, the emissions in the Off-Design operational 211

mode will also be calculated using linear proportionality with emissions at the nominal point of the C30HEV, increased by a 50% safety factor .Table 5 shows the limit values for emissions set by the Directive 98/69/CE [11]. Table 5 . EURO Directives on Emissions Normative EURO 5 EURO 4

NOX [g/km] 0,04 0,08

HC [g/km] 0,05 0,1

CO [g/km] 0,5 1

PM [g/km] 0,0125 0,025

Emissions for the LETHE@ City car configuration, on urban routes, are obtained by similar calculations, and are shown in table 6. Table 6. Emissions from the LETHE@ City Car on urban routes Pollutants NOx HC CO PM

Emissions [g/kWh] 5,81E-02 9,39E-02 9,17E-02 9,17E-04

Emissions [g/km] 6,18E-03 9,99E-03 9,75E-03 9,75E-05

Table 7 shows the percentage reductions of pollutant substances compared to the C30HEV and the EURO 5 Directive. Table 7. Percentage reduction of emissions of the LETHE@ City Car Pollutant C30HEV EURO 5 NOx HC CO PM

-70 % -70 % -70 % -70 %

-84,5 % -80,6 % -98 % -99,2 %

Such low emissions are justified by the intermittent use of the GT. On an urban mission during which 11 km are covered in 1959 seconds, the GT supplies power for only 440 seconds (including the time required to recharge the battery package at the end of the mission): therefore more than 75% of the mission is performed in electrical mode. The emissions are calculated on the entire route, considering both the electrical mode and the hybrid mode (GT switched on), and thanks to the net prevalence of electrical drive, we obtain the low emissions values displayed on Table 8. Table 8. Lethe@ data Consumption Pollutants emissions [km/l] [l/100 km] [g/kWh] NOx -80% 28 3.5 171 HC -80% CO -90% PM -90%

212

Table 9 : Capstone C30HEV (diesel fuel) emissions [15] at full load and EU limits Emissions units

EURO 5 (2008)

EURO 6 (2013) g/kWh

Capstone C30HEV

2.00 1.50 0.46

0.50 1.50 0.13

0.60 1.17 0.004

NOx CO PM

Table 10. LETHE® City Car emissions on urban routes and comparison to a commercial city car (SBZ srl 2011 data) LETHE® city car

pollutants

Smart*

Emissions [g/km]

NOx

0.00618

0.06

HC

0.00999

n.a.

CO

0.00975

EURO V

PM

0.0000975

EURO V

CO2

75

100

Table 11: Total emissions of a reference fleet composed by 500 gasoline vehicle, with individual annual mileage of 10000 km and comparison with a Lethe® fleet composed by 500 City cars (same mileage) Total NOx total CO2, t/year PM10Total particulate, Actual commercial fleet emission, kg/year kg/year Passenger sedan, gasoline 300 510 n.a. TOTAL EMISSIONS 300 510 n.a. PM10Total particulate, Total NOx total CO , t/year 2 LETHE® fleet kg/year emission, kg/year City Car 30.9 375 450 TOTAL EMISSIONS 30.9 375 450 The analysis of table 9 indicates that the Capstone 30HEV has lower emissions in comparison with those defined by European directives (in fact, it would satisfy the EURO 6). The comparison with other current commercial vehicles (tables 10 and 11) is also positive, thanks to the peculiarity of the operational mode of the turbogas group, that is not designed to supply power to the drive train, but to recharge the battery package (range extender) or to complement the battery power during power surges.

9. Conclusions The technical and economic feasibility of our design for a hybrid passenger sedan “LETHE” has been positively evaluated in previous papers [3,4]. The most important innovation in this project are the advantages offered by the adoption of a GT in lieu of the traditional thermal motor (ICE). This is not a complete “revolution” in the concept of cars as we know it today, but rather a simple reorganisation of the components. The energy flows management logic (VMU) for a gas-turbinedriven hybrid propulsion system has been previously described in detail [2]: it provides proper operational mode under all driving conditions. The application to possible configurations has been studied, the configurations being differentiated by the presence or absence of the dynamic storage 213

unit (the flywheel) and by different battery recharge modes. All simulations confirm that the LETHE vehicle is a competitive solution with respect to traditional ICE vehicles and also to other current hybrid vehicles: the calculated fuel consumptions is 29 km/l for urban cycles (compared respectively on 20 km/l for a current diesel vehicle, and 16 km/l for a gasoline sedan). All these advantages result in remarkable advantages in terms of weight, size and furthermore offer the opportunity of using a multi-fuel thermal engine (the GT) that can work with all types of liquid and gaseous fuels currently available on the market, thus reducing the economic effects of fuel price fluctuations. The analysis carried out in this article is synthetically expressed in Table 12, which shows the consumption and emissions of the LETHE hybrid according to the described procedure. This table contains data obtained by previous simulations and by the present assumptions [3, 4]: it shows a consumption reductions of about 30%, for a hybrid city car powered by methane compared to current commercial vehicles. With regard to emissions, we have highlighted the drastic reduction in all main pollutants emitted from the thermal engine compared to the values prescribed by current regulations, made possible by the optimisation of a thermal motor that operates mostly at design point. It must be underscored that it is the adoption of a Hybrid Series configuration (HS) that makes the use of GT device possible. In a global context of a simultaneous reduction of greenhouse gas emissions, of the consumption of fossil fuels, and of city pollution, it is clear that the benefits introduced by the HS vehicle would provide an immediate response to even the most stringent environmental regulations.

Nomenclature BP Battery Package BRC Breaking Recovery Coefficient BRL Battery Recharge Limit C Type of battery recharge limit DOD Depth of Charge EM Electric Motor GT Gas Turbine set GTHV Gas Turbine Hybrid Vehicle HD Hybridization Degree HS Hybrid Series HV Hybrid Vehicle ICE Internal Combustion Engine IGV Inlet Guided Vanes LETHE Low Emission Turbo Hybrid Electric Vehicle MTBF Mean Time Before Failure PM Particulate SOC State of Charge UDR1 University of Roma 1 VMU Vehicle Management Unit

References [1]

R. Capata et al.: A Gas Turbine-Based Hybrid Vehicle-Part II: technological and configuration issues, JERT , V. 125 n. 2, July 2003, 777-782

214

[2]

[3] [4]

[5]

[6]

[7] [8] [9] [10] [11] [12]

[13] [14] [15] [16] [17]

R. Capata, M.Lora. The Comparative assessment and selection of an “optimal” configuration for a Gas Turbine-Based Hybrid city car, JERT.2008, vol. 129, n.2, pagg. 107-117 R. Capata, A. Coccia: Procedure for the design of a Hybrid-Series vehicle at UDR1 and the Hybridization Degree choice. Energies 2010, vol. 3 , pagg. 450-461 R. Capata, A. Coccia, M. Lora: A proposal for the CO2 abatement in urban areas: the UDR1–Lethe© turbo-hybrid vehicle. Energies, Hybrid Vehicle Special Issue – Energies 2011, 4(3), 368-388; doi:10.3390/en4030368 F. Ciaralli: Evaluation and determination of efficiency index in the Combined and TurboGas Power Plant. Master Degree Thesis (In Italian) Dept. Of Mechanical and Aerospace Engineering, University of Roma 1 “Sapienza”, Rome, 2001 E. Cioffarelli, E Sciubba: A new type of gas turbine based-hybrid propulsion systemPart I: concept development, definition of mission parameters and preliminary sizing, Proc. AES/ASME Winter Meeting, Orlando, FL, USA, 2000. Minotti A., E. Sciubba: Comparison of LES calculation and experimental data for an ultra micro combustion chamber G. Pede, ENEA (National Agency of Energy and Environment), Personal Communication, 2009 Italian Automobile Club. Available at: [accessed 12.3.2012] Audi official web site. Available at [accessed 12.12.2011] Italian gas Company web site. Available at: www.autogasitalia.it [accessed 02.02.2012] Capstone web site. Available at: www.interstatepower.us/Capstone/Document/Library/Application/Guides/480009_HEV _Application_Guide.pdf [accessed 11.11.2011] Italian Automotive Magazine web site. Available at: [accessed 01.12.2011] Italian Fuel Prize official website. Available at: [accessed 01.12.2011] SBZ SrL, Personal Communication, 2011 Ansaldo website, Products list. Available at: [accessed 14.10.2011] Burke AF: Cycle Life Considerations for Batteries in Electric and Hybrid Vehicles, SAE Technical Paper Series #951951, reprinted in Electric and Hybrid VehiclesImplementation of Technology (SP-1105); Future Transportation Technology Conference and Exposition, Costa Mesa, CA. August, 1995. Publication No. UCD-ITSRP-95-21

215

A variational optimization of a finite-time thermal cycle with a Stefan-Boltzmann heat transfer law J.C.Chimal-Eguiaa, N.Sanchez-Salasb a Centro de Investigación en Computación del IPN, México D.F, México, [email protected] b Escuela Superior de Física y Matemáticas del IPN, México D.F., México, [email protected],

TL

TR

TR

TH

TR

TL

nT n

1

T n (n 1)

T2

TR

T TR

TH

T

Th

y TL Tl

TH

P ( x, y )

x TH

TH

TRn

TR

TH

0

0

TR

T

Tl

1

TL TH

1

CA

P

TR

TL

TR

Tl Th

Th

TL

4T 5 3 T 4 TR

TH 5 h

T

TR4

0

T

Tl

Th 4 h

4 H

4T

3 T

T

0

4Tl 5

3 Tl 4

TL4

0

Th

Tl

Tl

TL

Tl 1

2

5 h

3T T

)5

3Tl 4 (1

4T

TH4 (1

4 l h

4 l H

TT

0

) 4Tl 4

0 0.2029

CA

TL TH

1

L W

SW

SW 0

q dt T

0

q

q

(T

n R

n

T )

TR L

(TRn

L

T n)

0

L

TR g (T , TR , TL , )

(TRn

T n) T

0

dt

Tf (T , TR , TL , ) dt

0

0

ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

216

MODELING AND SIMULATION OF A BOILER UNIT FOR STEAM POWER PLANTS L.Moliterno, C.Toro University of Rome “Sapienza” , Italy

INTRODUCTION The modeling and steady state simulation of large-size boiler units for steam power plants have been performed. The component “boiler” has been developed and implemented into the library of a modular object-oriented process simulator, CAMEL-Pro™ to perform the simulation and the exergy analysis of the single component and of the global steam power plant. The model created is than applied to a large size steam generator for a thermal power plant of 320 MW, powered by oil and with a reheat of the steam. MODEL The model implemented is based on mass and energy balances and proper equations for the evaluation of the heat exchange coefficients are introduced. Energy balances are performed not only for the whole plant but also for each component in order to evaluate the distribution of the thermodynamic inefficiencies throughout the system. The equation system has been written using CAMEL-Pro™ Simulation Software which is an “Object oriented” modular code (C++) equipped with a WINDOWS-compatible and user friendly graphical user interface (C#)

Scheme of the model

Unknowns • Air/gas stream 12 unknowns (8 stream)

𝑚, 𝑝, 𝑇, ℎ, 𝑠, 𝑒𝑥, 𝑞, 𝑥𝑂2 , 𝑥𝑁2 , 𝑥𝐶𝑂2 , 𝑥𝐻2𝑂 , 𝑣 • Steam/water stream 7 unknowns (9 stream)

𝑚, 𝑝, 𝑇, ℎ, 𝑠, 𝑒𝑥, 𝑞 • Fuel 5 unknowns

Heat transfer The heat transferred in the furnace to the vaporizer’s tube banks has been calculated as a percentage of the heat introduced with fuel (Orrok’s method). The emissivity of the gas necessary to calculate the radiation heat transfer has been obtained with the equation proposed from Ganapathy.

Characteristics of the 𝑚, 𝐿𝐻𝑉, 𝑥𝐶 , (𝑥𝐻), 𝛼 generator: 320 MW nominal power; The problem presents 165 •Oil fuel; unknowns, 128 equation has been •Forced circulation of water; written, so to perform the simulation •Pusher fan; 37 boundary conditions are necessary.

Legend : — Air/Gas; — Steam/Water; — Fuel;

RESULTS The geometric and operative data are referred to the ENEL’s power plant of Sermide Boundary condition

Steam Power Plant Simulation With CAMEL-Pro™, using the component “boiler” developed, has been simulated a steam power plant high pressure side. The imagines below represent s the plant’s layout in CAMEL-Pro™

Comparison between simulation results and ENEL’s power plant of Sermide data

References • Donatello Annaratone, “Steam Generator - Description and Design” 2008 SpringerVerlag Berlin Heidelberg • V.Ganapathy, “Industrial boilers and heat recovery steam generators” 2003 Marcel Dekker, Inc. •http://www.turbomachinery.it, CAMEL-Pro™ Manual V.4.0

ECOS 2012 - the 25th International Conference on Efficiency, Cost, Optimization, Simulation and Environmental Impact of Energy Systems JUNE 26-29, 2012, PERUGIA, ITALY

217

NUMERICAL MODELLING OF STRAW COMBUSTION IN A MOVING BED COMBUSTOR Biljana Miljkovi , Ivan Pešenjanski, Borivoj Stepanov, Vladimir Milosavljevi , Vladimir Rajs Faculty of Technical Sciences, Novi Sad, Serbia, [email protected]

INTRODUCTION and BECKGROUND Combustion of wheat straw in Serbia is a perspective way of energy conversion but devices for agricultural waste combustion are still in developing phase and good enough design solution still does not exist. Since the combustion of straw for power generation is a relatively new concept, the design and operating conditions are not fully optimized. That is why grate-fired boilers burning straw are often associated with high emission levels and relatively poor fuel burnout. Mathematical modelling thereby becomes a cost-effective alternative to exhaustive testing in designing, retrofitting, analyzing and optimizing the performance of combustion systems. A moving grate furnace is a typical way to burn solid biomass in many combustion plants and a furnace is the simplest combustion technology. Despite its apparent simplicity, direct combustion is an extremely complex process from a technological point of view. Bed height

5

3

4

7 9 2 1 8 M

During the continual combustion process that happens in such hypothetical moving bed, moisture evaporation front, straw pyrolysis and fuel combustion front are formed, as illustrated in figure 2.

6

Fig 1. Real combustion equipment for bale straw

0.4 m

T=1500 K I III

IV

II

Straw

1m

Bed length

Primary air

Fig 2. Contour of fuel by combustion in a horizontal bed I-raw straw, II-dry straw, III-pyrolysis front, IV-combustion front

RESULTS and DISCUSSION

Fig 3.a) Simulated results of bed profile, b) temperature profile of solid phase, c) volatiles profile, d) char profile and e) pressure profile in the bed

1. As shown in figure 3, during combustion process in a moving bed, two stages are defined, (I) the preheat stage at the start of the grate, where the solid temperature of the surface layer rises sharply from room level to a peak temperature (about 550 K) and (II) intense reaction stage after the heating stage, where the top bed temperature starts to rise and reaches about 700 K. 2. In the preheat stage (I) drying process is finished and partially devolatilisation process, but there are no intense combustion reactions and hence no visible degradation of the bed. The heating zone in this stage stays very thin, only a few millimeters. 3. In the intense reaction stage (II), after heating, drying and devolatilisation processes, the local bed temperature rises to a peak value at which point combustion process of volatiles starts, followed by combustion of char. The heat from the char oxidation and other combustible material remains in the bed, increasing the bed temperature to a higher level than in the heating stage. CONCLUSIONS An original mathematical model is developed and simulations are carried out for the combustion of straw in a moving bed. Model includes: moisture evaporation, straw pyrolysis, volatiles and char combustion. The model is able to predict the influence of main processes occurring in straw bed combustion. Mathematical modelling provides detailed information on the burning processes, which is otherwise very difficult to obtain by using conventional experimental techniques. Very important result of simulations is the time needed for complete burnout of the bed. Thus, the model can be used to perform investigations of different furnace conditions and may be used for the optimisation of existing furnaces and development of new ones. ACKNOWLEDGMENTS This work steams from our ongoing investigation “Development of methods, sensors and systems for monitoring of water, air and soil quality”, funded by the Serbian Ministry of Science (Project No. III 43008) ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

218

PHYSICOCHEMICAL EVALUATION OF THE PROPIERTIES OF THE COKE FORMED AT RADIATION AREA OF LIGHT HYDROCARBONS PYROLYSIS FURNACE IN PETROCHEMICAL INDUSTRY

Figure 7. Micrographs layer 2- section E 1,2

Jaqueline Saavedra Ruedaa, Angélica María Carreño Parrab, María del Rosario Pérez Trejosc, Dionisio Laverde Catañod, Diego Bonilla Duartee, Jorge Leonardo Rodríguez Jiménezf, Laura María Díaz Burgosg aInstituto

Colombiano del petróleo, Ecopetrol, Santander, Colombia, [email protected] bInstituto Colombiano del petróleo, Ecopetrol, Santander, Colombia, angelica.carreñ[email protected] cInstituto Colombiano del petróleo, Ecopetrol, Santander, Colombia, [email protected] dDepartamento de Ingeniería Química, Universidad Industrial de Santander, Colombia, [email protected] eDepartamento de Ingeniería Química, Universidad Industrial de Santander, Colombia, [email protected] fDepartamento de Ingeniería Química, Universidad Industrial de Santander, Colombia, [email protected] gUT-TIP PETROLABIN- Santander, Colombia, [email protected]

Real Density and porosity of coke

INTRODUCTION

Property Real Density[g/cm3] Apparent Density [g/cm3] Pore volume [cm3/g] Porosity

Figure 2. Coils of the radiation spot Source: AutoCAD-2010. Amezquita J.C

Figure 1. Ethane pyrolysis furnace for the production of ethylene.

The porosity, influences the mechanical and physical properties of coke, because the existing pores generate spaces between its molecules. Furthermore, the porosity affeds the the kinetics of reactions, particularly in the combustion and gasification.

Valor Promedio 1,96 ± 0,01 1,01 ± 0,01 0,24 ± 0,13 31,63 ± 0,14

Table 2. Parameters calculated from the skeletal density

Figure 3. Coke formed in the inner walls of the tubes.

Different studies using physicochemical characterization had been focused to cokes obtained from laboratories, under controlled conditions and with magnitudes relatively small, respect to the industrial scale, what makes it more easy and predictable the analysis of heat and mass transference phenomena.

Coke Chemical Composition Substance [%w/w] Fixed Carbon 95,68 Ash 0,37 Moisture residual 0,03 Volatile matter 3,92 Table 3. Coque proximate Analysis

Elemental Analysis Substance [%p/p] Total sulfur 50 m, overall height: 90 to 100 m, output: 2 to 3 MW Technip partner of the Iberdrola – Eole-RES Consortium awarded 500 MW Saint-Brieuc Offshore Wind Project, France Ethanol Hundreds of references globally A know-how based on the acquisition of Speichim in 2000 Proprietary technologies

Diesel Close cooperation with Axens on a unique process (Esterfip-H) Collaboration with Neste Oil for various project using their NexBtL proprietary technology

New Generation of biofuels: Ligno-cellulosic ethanol; BtL; Others… EUROPEAN PROJECT CACHET II LED BY BP 8 partners overall including Technip (leader in the design of reactor) Funded by EU commission under the FP7 The project is about an Hydrogen (H2) permeable membrane reactors for precombustion carbon dioxide capture in both coal and gas fired power stations. Main benefits are : Combination of efficient conversion of syngas into hydrogen fuel with capture of the remaining carbon dioxide in one reactor. The carbon dioxide is produced at high pressure,reducing the compression energy for transport and storage. ECOS 2012 - THE 25TH INTERNATIONAL CONFERENCE ON EFFICIENCY, COST, OPTIMIZATION, SIMULATION AND ENVIRONMENTAL IMPACT OF ENERGY SYSTEMS JUNE 26-29, 2012, PERUGIA, ITALY

222