Understanding and bridging the energy performance gap in building ...

Available online at www.sciencedirect.com

ScienceDirect Energy Procedia 00 (2017) 000–000 www.elsevier.com/locate/procedia

CISBAT 2017 International Conference – Future Buildings & Districts – Energy Efficiency from Nano to Urban Scale, CISBAT 2017 6-8 September 2017, Lausanne, Switzerland

Understanding and bridging the energy performance gap in building retrofit Jad Khoury*, Zeinab Alameddine, Pierre Hollmuller University of Geneva, Energy System Group, Institute for Environmental Sciences & Department F.-A. Forel and aquatic science, Faculty of Science, 1205 Geneva, Switzerland

Abstract On the basis of a representative set of examples, we tackle following issues concerning the energy performance gap which is commonly observed in retrofit of multifamily buildings: quantification of the main factors behind the gap, in relation with simulation; decomposition of the gap, by combination of the most sensitive parameters, in relation with observed values; exploration of how to bridge the gap, on the basis of two best practice examples. The results show that the gap consists of: (i) the difference between optimal yet realistic conditions of use, and standard values used in the simulation process; (ii) the difference between actual and optimal conditions of use, which corresponds to the optimization potential. It proves that significant reduction of the performance gap is possible through building optimization and responsible behavior. © 2017 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of the scientific committee of the scientific committee of the CISBAT 2017 International Conference – Future Buildings & Districts – Energy Efficiency from Nano to Urban Scale. Keywords: energy performance gap, building retrofit, case studies, sensitivity analysis, optimization measures

* Corresponding author. Tel.: +41 22 379 0277 E-mail address: [email protected], [email protected] 1876-6102 © 2017 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of the scientific committee of the CISBAT 2017 International Conference – Future Buildings & Districts – Energy Efficiency from Nano to Urban Scale.

2

KHOURY Jad, ALAMEDDINE Zeinab, HOLLMULLER Pierre / Energy Procedia 00 (2017) 000–000

1. Introduction Achieving the energy savings targets set in the Swiss Energy Strategy 2050 (ES2050) requires a reduction of about 45% in energy use for space heating in buildings by 2035, and a 64% reduction by 2050, compared to 2010 levels. Given the slow transformation of the building stock, particular emphasis should be placed on improving the quantity and quality of building energy retrofits. However, current research indicates a serious performance gap between theoretical and actual energy savings in real-life condition of execution, operation and use [1,2]. According to a recent study on a representative sample of retrofitted multifamily buildings, the achieved fractions of the theoretical savings for space heating are in the range of 30% to 65% [3,4]. It also shows that, under current practices (without optimization), about half of the theoretical saving potential of the Swiss building stock could be achieved in reality, making it difficult to achieve the ambitious goals of the ES2050. In the light of these findings, the purpose of the present contribution is threefold:  to examine the determinant factors behind the energy performance gap and to quantify their relative importance, via numerical sensitivity analysis;  to explain and decompose the performance gap, by combination of the most sensitive simulation parameters, in relation with observed values;  to explore how to bridge the performance gap, on the basis of two best practice examples. Nomenclature Qh Qh, norm

space heating demand space heating demand, in standard / normed condition of use

2. Methodology 26 retrofit operations of residential multifamily buildings were selected from the retrofit building permit requests which were submitted to the Energy Office of Canton Geneva, between 2004 and 2012. The different stages of the selection process are detailed in [4]. Of these, 20 retrofit case studies had a detailed energy balance in the building permit (design values calculated by the engineers) and are therefore considered for following analysis. These cases involve 94 alleys (building entrances), comprising about 2 225 dwellings and covering a total energy reference area of approximately 210 000 m2. The buildings are fairly representative of those constructed during the post-war period (1946-1980) in the canton of Geneva [4] and more generally in Switzerland, and also offer the most important energy saving potential. In a first step, a sensitivity analysis of the simulated space heating demand is performed separately for 10 input parameters and over 20 retrofit case studies, using a certified thermal calculation software. The analysis consists in the identification of the most determining factors behind the gap and the quantification of their relative importance on the space heating demand calculation. In a second step, the combination of the most sensitive parameters allows to understand the observed discrepancy between theoretical and actual energy savings. In a third step, some mitigation measures capable of reducing the observed performance gap were implemented on two deep retrofit case studies and the respective reduction potential was assessed. 3. Main factors behind the gap There are several independent and interdependent factors which can explain why retrofitted buildings do not perform as well as predicted. These factors can occur at different stages of the building retrofit process. Three parameter sets are identified in this study as the main causes for the discrepancy observed: i) inaccuracies due to the use of SIA [5] standard values in the method to calculate the theoretical savings; ii) quality of other input data used for the simulation that depends on the choices made by the model operator (e.g. design weather data, regulation type, shading factor, calculated surfaces, etc.), or model limitations; iii) other factors related to quality of execution, operation, monitoring

J. Khoury et al. / Energy Procedia 00 (2017) 000–000

3

and user behavior (both occupant and energy operator). In this section, a sensitivity analysis is performed separately for 10 simulation input parameters of set (i) and (ii), for 20 retrofit case studies (B1-B20). 3.1. Inaccuracies due to the use of standard values The sensitivity of Qh to the use of standard SIA values is analyzed separately for the 6 corresponding parameters (Tab. 1), for the 20 buildings. The analysis shows that the two most sensitive parameters are the indoor temperature and the air flow rates, which play a major role in the space heating demand calculation (Qh). Table 1. Sensitivity of simulated space heating demand (Qh) to the use of standard values. Parameter (set i)

Unit

Indoor temperature

°C

20

23

+1

10.7% ± 0.9%

Fresh air flow rates

m³/h.m²

0.7

1-1.5*

+0.2

17.5 ± 0.7 MJ/m²

Utilization period

h

12

-

+2 -2

-2.4 ± 0.3 MJ/m² 2.4 ± 0.3 MJ/m²

Electricity factor reduction

-

0.7

0.8

+0.1 -0.1

-5.1 ± 0.7 MJ/m² 5.4 ± 0.7 MJ/m²

Annual electricity consumption

MJ/m²

100

124

300

+25 -25

-8.6 ± 1.2 MJ/m² 9.4 ± 1.2 MJ/m²

250

Occupancy

m²/P

+10 -10

2.1 ± 0.3 MJ/m² -7.0 ± 0.9 MJ/m²

400

Sample average

Action

400

350

40

53

Qh [MJ/m2]

300

200

250

Impact on Qh, norm (avg±std)**

* Mechanical ventilation: 1 m³/h.m² (average of the measured200values during the winter heating period); 150 Window opening: 0.5 m³/h.m² (estimated value). ** avg: average value; std: standard deviation. 100

150 100 50

50

0 0

B20 B8 B7 B14 B3 B17 B9 B15 B2 B10 B16 B19 B18 B1 B5 B6 B11 B12 B13 B4


Indoor temperature Buildings

T_21

T_22

T_23

Air flow rates

Buildings

T_24

T_25

Base

400

400

350

350

300

300

Qh [MJ/m 2]

Qh [MJ/m 2]

Base

250 200 150

V_0.9

V_1.1

V_1.3

V_1.5

V_1.7

250

200 150

100

100

50

50

0

0 B20 B8 B7 B14 B3 B17 B9 B15 B2 B10 B16 B19 B18 B1 B5 B6 B11 B12 B13 B4


Buildings

25

25%

20 15 10 5 0

Buildings 30

20% 15% 10% 5%

100

200

Qh, norm [MJ/m²]

300

25 20

15 10 5 0

0% 0

30%

∆Qh, +0.2 m³/m²/h [%]

30%

∆Qh, +0.2 m³/m²/h [MJ/m²]

30

∆Qh, Tint +1 C [%]

∆Qh, Tint +1 C [MJ/m²]

Qh [MJ/m2]

350

Standard value

0

100

200

300

Qh, norm [MJ/m²]

Indoor temperature +1 C  Increase of Qh by 11%

1%

25% 20% 15% 10%

5% 0%

0

100

200

Qh, norm [MJ/m²]

300

0

100

200

Qh, norm [MJ/m²]

Air flow rates +0.2 m³/m²/h  Increase of Qh by 17-18 MJ/m²

Fig. 1. Sensitivity of simulated space heating demand (Qh) to various indoor temperature and airflow rates (results for 20 retrofit case studies B1- B20 sorted by descending order of heat losses)

300

4


It appears that each °C increase in the indoor temperature leads to a raise of the normed space heating demand (Qh, norm) by around 11%, whereas an elevation of the air flow rate of 0.2 m3/h.m2 leads to an increase of about 17.5 MJ/m2 (Fig. 1). Note that the actual average indoor temperature (23°C) measured in 6 retrofitted buildings during the heating period greatly exceeds the standard SIA value of 20°C. Similarly, the measured direct air flow rates amounts to 1 m3/h.m2 on average, and to 1.5 m3/h.m2 if we consider indirect air flow resulting from window openings, compared to the standard SIA value of 0.7 m3/h.m2. 3.2. Quality of the input data used for the simulation The choices made by the model operator in simulating Qh can also contribute to the discrepancy between design and actual performances. In this study, the analysis performed separately on 4 parameters (Tab. 2) shows that, depending on the user choice or appraisal error, uncertainties on the simulated space heating demand can be in the order of 5 – 10%. Note that the SIA 380/1 [5] does not provide standard values for these parameters. Table 2. Sensitivity of simulated space heating demand (Qh) to choices or possible errors made by the model operator. Parameter (set ii) Weather data (for Geneva-Cointrin) SIA regulation type (0K, 1K, 2K)*

Action

Impact on Qh,norm (avg±std)**

SIA 2028 instead of SIA 381/3

-5.4% ± 1.5%

+1

10.2% ± 2.4%

Shading factor (25%, 35%, 45%)

+10%

4.3% ± 1.9%

Energy reference area

+10% -10%

-7.7% ± 0.6% 9.5% ± 0.7%

* Increase in the indoor temperature for non-performing regulation. ** avg: average value; std: standard deviation.

3.3. Other parameters The discrepancy between design and actual performances can further arise from factors which can occur during the execution and operation phase of the retrofit process. Factors like poor commissioning, design changes during execution phase, malfunctioning of technical systems were not assessed in this study, but are partially embedded in the following section. 4. Decomposition of the performance gap In a previous study [3] we analyzed the relation between the theoretical and actual energy savings for space heating on 10 of these buildings (B1-B10), for which actual final energy demand before and after retrofit was available. In each case, actual values of Qh were derived by taking into account average heating system efficiencies (depending on the energy career) and an average heat demand for domestic hot water (as derived from a benchmark). Theoretical savings are defined by the difference between the actual Qh before retrofit and the expected Qh after retrofit, as stated in the building permit and calculated under standard conditions of use; actual savings are defined by the difference between the actual values Qh before and after retrofit. As shown in Fig. 2 (blue points), the achieved fractions of the theoretical savings for Qh are in the range of 30% to 65%. In this section, we try to analyze this performance gap by simulating different scenarios, with several combination of the most sensitive parameters identified above:  V0: reference scenario according to SIA standard condition of use.  V1: real data for occupancy, annual electricity consumption and electricity factor reduction.  V3: optimized scenario, with V1 + an indoor temperature of 21 and an air flow rate of 1.1 m³/m²/h.  V4 – V5: current practice scenario for the existing building stock (without optimization), with V1 + an indoor temperature of 23 – 24°C and an air flow rates of 1.3 – 1.5 m3/h.m2.

J. Khoury et al. / Energy Procedia 00 (2017) 000–000

5

This analysis is done on 4 of the 10 buildings (B1, B4, B5, B9), representing the observed range of theoretical savings. As a result, the performance gap can be divided into two parts: (i) the potential for optimization, which corresponds to the difference between the actual conditions of use (V4-V5) and the optimal ones (V3); (ii) the Figure forthe CISBAT 2017 (V0) and optimal conditions of use (V3). difference between standard 600

80%

70%

50% 40% 30% 20% 10%

Actual energy savings [MJ/m²/yr]

500 60%

① V3

V4

V5

400

② V0

V1

B1 B2

300

200

B3

B4

B6 B5

100 B8

0%

① Potential for optimization ② Difference between standard

B7

B9

B10

and optimal conditions of use

0 0

100

200

300

400

500

600

700

Theoretical energy savings SH [MJ/m²/yr]

Fig. 2. Decomposition of the performance gap in building retrofit.

5. Successful examples of reducing the gap

600


The last part of this study explores possibilities of how to bridge the performance gap, on the basis of two 500 further examples (B11, B12-B13), which reflect today’s best practices (Fig. 3). After retrofit, these buildings perform better than the trend observed in Fig. 2, due to the particular attention paid during the design and execution phase (dark blue 400 points); furthermore, a series of optimization measures were conducted after the retrofit process (red points).

B11


Minergie retrofit B13 (2010-2011) B12

Minergie-P retrofit (2013-2014)

B11 B13 B12

Increasing actual energy savings through building optimization ①

600 300

500 200

100 400

B10

0 300 0

100

200

100

Rest: difference between actual and standard conditions of use ②

B

0

0

Fig. 3. Examples of optimization after retrofit (B11, B12-B13).

100

6


Improvements of building B11 (Minergie retrofit) includes correction of the dysfunctions of the solar thermal system, gradual reduction of the indoor temperature using a model predictive control (target: 21.5°C), hydraulic balancing of the heating system and an adjustment of the heating curve (further information in [6]). For buildings B12 and B13 (Minergie-P retrofit), the implemented measures were an adjustment of the mixing valves of the heating system and of the heating curve, as well as adapting the settings of the heat recovery ventilation system, which was unduly bypassed, resulting in a significant increase of the heat recovery efficiency during the mid-season (from 30% before, up to 80% after optimization [7]). After optimization, the achieved fraction of the theoretical saving potential has in both cases been increased from 65% to around 80%. The rest is mainly due to the difference between the standard and optimal conditions of use, in particular the use of optimistic/inaccurate values of the SIA standard for calculating the theoretical savings. 6. Conclusions This study focuses on a previously analyzed representative sample of retrofitted multifamily buildings, for which the achieved fractions of the theoretical savings for space heating are in the range of 30% to 65%. A numerical sensitivity analysis concerning standard parameters points out the particular influence of the indoor temperature and the air flow rates (each additional degree leads to an increase of the space heating demand by around 11%, whereas an elevation of the air flow rate of 0.2 m3/h.m2 leads to an increase of about 17 MJ/m2). The performance gap of the building sample is analyzed by way of different scenarios, with several combination of the most sensitive parameters. It shows that the performance gap can be divided into two parts: (i) the potential for optimization, which corresponds to the difference between the actual and the optimal conditions of use; (ii) the difference between the standard and the optimal conditions of use. While the main goal of using standard calculation method (in normed condition of use) is to compare the calculated energy use of buildings with the limit and target values, on a common basis, the question remains whether it makes sense to use such a method to estimate real energy savings at building level, or to design thermal retrofit strategies and policies. Finally, the last part of the study explores possibilities of how to bridge the performance gap, on the basis of two best practice examples. The result shows that the optimization measures conducted after the retrofit process allows to increase in both cases the achieved fraction of the theoretical saving potential from 65% to around 80%. This demonstrated that a significant reduction of the performance gap is possible through building optimization and responsible behavior (both occupant and energy operator). Acknowledgements The authors would like to thank the authorities of Canton Geneva (OAC, OCEN, OCSTAT) for the provision of the data behind this study. The study was funded by SFOE and OCEN in the frame of the Compare-Renove project [8], as well as by CTI in the frame of the SCCER FEEB&D project. References [1] BFE (2016). Erfolgskontrolle Gebäudeenergiestandards 2014-2015, Bundesamt für Energie BFE, Bern. [2] Majcen, D. (2016). Predicting energy consumption and savings in the housing stock: A performance gap analysis in the Netherlands, PhD thesis, Delft University of Technology, OTB - Research for the Built Environment. [3] Khoury J., Hollmuller P., Lachal B. (2016). Energy performance gap in building retrofit : characterization and effect on the energy saving potential. In: 19. Status-Seminar «Forschen für den Bau im Kontext von Energie und Umwelt». Zurich, 8-9 September 2016. [4] Khoury, J. (2014). Rénovation énergétique des bâtiments résidentiels collectifs: état des lieux, retours d’expérience et potentiels du parc genevois, Thèse de doctorat, Université de Genève. [5] SIA (2009). SIA 380/1: L’énergie thermique dans le bâtiment. [6] Flourentzou F., Pantet S. Raisons et remèdes da la surconsommation des bâtiments locatifs après rénovation. In: 19. Status-Seminar «Forschen für den Bau im Kontext von Energie und Umwelt». Zurich, 8-9 September 2016. [7] Tornare, G. et al. (2016). Rapport technique et de communication du projet d’assainissement Minergie-P des immeubles « La Cigale » (GE) – Chauffage par pompes à chaleur solaires couplées à des stocks à changement de phase. Research project SFOE. [8] Khoury, J. et al. (Final report in preparation). COMPARE-RENOVE – Du catalogue de solutions à la performance réelle des rénovations énergétiques. Research project SFOE 2013-2016. Available at: https://www.unige.ch/energie/fr/activites/axes/efficacite/comparenove/