Effect of Granule Sizes on the Performance of Upflow Anaerobic ...

Available online at www.sciencedirect.com

ScienceDirect Energy Procedia 79 (2015) 90 – 97

2015 International Conference on Alternative Energy in Developing Countries and Emerging Economies

Effect of Granule Sizes on the Performance of Upflow Anaerobic Sludge Blanket (UASB) Reactors for Cassava Wastewater Treatment Sunwanee Jijaia,b, Galaya Srisuwana, Sompong O-thongc,d, Norli Ismaile and Chairat Siripatanaa,b* a

School of Engineering and Resources, Walailak University, Nakhon Si Thammarat 80161, Thailand b Renewable Energy Research unit, Walailak University, Nakhon Si Thammarat 80161, Thailand c Department of Biology, Faculty of Science, Thaksin University, Patthalung 93110, Thailand d Research Center in Energy and Environment, Thaksin University, Phatthalung 93110, Thailand e School of Industrial Technology, Universiti Sains Malaysia, Penang, 11800, Malaysia

Abstract The laboratory-scale UASB reactors were operated at five different hydraulic retention times (HRTs). The various sizes of granules from three different sources: a cassava factory (CS), a seafood factory (SS), and a palm oil mill (PS), having the size range of 1.5-1.7 mm, 0.7-1.0 mm and 0.1-0.2 mm. respectively, were used as inocula for anaerobic digestion of cassava wastewater. For comparison, the first reactor with only granules from its own source (R1, CS) was treated as control. The other two reactors were inoculated with mixed granules from different sources (R2, CS+SS and R3, CS+PS). As HRT decreased from 5 days to 1 day, the organic removal efficiencies decreased from 91.49 to 43.23 %, 89.36 to 45.13 % and 87.23 to 32.69 % for R1, R2 and R3 respectively (or inversely with increasing OLR). In this study selected mathematical models including Monod, Contois, Grau second-order and Modified Stover-Kicannon kinetic models were applied to determine the substrate removal kinetics of UASB reactors. Kinetic parameters were determined through linear regression using experimental data obtained from the steady-state experiments and subsequently used to predict effluent COD. The results showed that Grau second-order and Modified Stover-Kicannon kinetic models were more suitable than the others for predicting the substrate removal for all different sizes of granules. In addition the Upflow Anaerobic Sludge Blanket (UASB) reactor with only granules from a cassava factory gave the best performance. © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license © 2015 The Authors. Published by Elsevier Ltd. (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the Organizing Committee of 2015 AEDCEE. Peer-review under responsibility of the Organizing Committee of 2015 AEDCEE

Keywords: UASB, Cassava wastewater, Granule sizes, Monod, Contois, Grau second-order, Modified Stover-Kicannon

* Corresponding author. Tel.: +6-675-672-309; fax: +6-675-672-399. E-mail address: [email protected].

1876-6102 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the Organizing Committee of 2015 AEDCEE doi:10.1016/j.egypro.2015.11.482

Sunwanee Jijai et al. / Energy Procedia 79 (2015) 90 – 97

1. Introduction There is a large potential to convert wastewaters from agro-industry into energy. In Thailand, a 'Strategic Plan for Renewable Energy Development' has been established since 2003. It aims to increase the share of renewable energy from 6.4% or 4,237 kilo tons of crude oil equivalent (ktoe) per year in 2008 to 20.3% of the commercial primary energy or 19,700 ktoe per year by the year 2022. Although Thailand is an agriculture country with the large volume of potential biogas feed stocks, only two major sources are currently utilized for biogas production: wastewaters from cassava starch factories and pig farms [1]. Agro-industries are major contributors to worldwide industrial pollution including Thailand. Effluents from many agro-industries are hazardous to the environment and require appropriate and comprehensive management approach [2]. The anaerobic digestion is an environmentally friendly to treat agro-industrial wastewater. It is widely used not only to treat wastewater but also to generating biogas, a clean and renewable energy which can substitute energy from fossil sources as targeted by current Thai energy policy. Many researchers found that wastewater from cassava starch factory are suitable for anaerobic process particularly for UASB reactor in which wastewater was supplied from the bottom of the reactor and the organic matter is digested as the wastewater moves upward. During digestion, methane gas bubbles are produced and carry the sludges upwards, resulting in the formation of dense sludge flocs (granules) that readily settle [3, 4]. The successful treatment in UASB reactor is principally attributed to the formation of anaerobic granules in sludge bed. The granule size is an important parameter directly influences the performance of reactor. Other factors affecting the performance of UASB reactor include temperature, organic loading rate, pH, alkalinity and nutrients etc [5]. The scope of this research is to study the effect of granule sizes on the performance of UASB rectors for cassava wastewater treatment in term of organic removal and biogas potential from different sizes of inocula/granules using kinetic models. The preliminary results in this work could be valuable for design and operation in continuous biogas plants. Nomenclature ܽ

ܵ଴ Τ‫ܭ‬௦ ή ܺ (per day)

ܾ

Constant for Grau second-order model (g VSS/l)

‫ܭ‬஻

Saturation value constant (g COD/l.day)

‫ܭ‬ௗ

The death rate constant (per day)

‫ܭ‬௦

Half saturation concentration (g COD/l)

݇௦

Grau second-order substrate removal rate constant (per day)

ܳ

Inflow rate (l/day)

ܴ௠௔௫

Maximum utilization rate constant (g COD/l.day)

ܵ଴ ǡ ܵ

Substrate concentration in the feed and effluent (g COD/l)

ܸ

The reactor volume (l)

ܺ଴ ǡ ܺǡ ܺ௘ Concentration of biomass in the feed, reactor and reactor effluent respectively (g VSS/l) ܻ

The yield coefficient (g VSS/ g COD)

ߤǡ ߤ௠௔௫ The specific growth rate and maximum specific growth rate, respectively (per day) Nomenclature (Cont’d) ߠ௖

Mean cell-residence time (day)

91

92


ߠு

Hydraulic retention time (day)

ߚ

Kinetic parameter (g COD/g biomass) for Contois model

2. Materials and methods 2.1 Wastewater and Seed The wastewater sample was collected from a cassava factory. Characteristics of wastewater are shown in Table 1. It was kept at 0-4 oC until used in the experiments. The granular sludges were collected from the methanogen fermentation stage of the UASB reactors from three sources: cassava factory (CS) having the size range of 1.5-1.7 mm, seafood factory (SS) having the size range of 0.7-1.0 mm and palm oil mill (PS) having the size range of 0.1-0.2 mm. The Specific methanogenic activities (SMA) were 0.28, 0.26 and 0.16 gCOD/gVSS.d respectively. Table 1. Basic parameter of cassava wastewater Parameter

pH

COD (g/l)

TKN (mg/l)

TP (mg/l)

TS (g/l)

VS (g/l)

SS (mg/l)

VSS (mg/l)

Value

5.0

18.8

320

70

16.3

15.5

1,900

250

Alkalinity (mg/l asCaCO3) 162.5

VFA (mg/l asCaCO3) 562.5

2.2 Reactor and Operating condition The three identical laboratory-scale UASB reactors were used in this study. They have cylindrical shape with 100 cm high, 5.4 cm internal diameter and 2.06 L working volumes. The feed was pump by peristaltic pump (Longer pump, Model BT 100-1F, DG-4 channel pump head) at the rate defined by HRT. Three reactors were operated continuously at five hydraulic retention times (HRTs) of 5, 4, 3, 2 and 1 day. The corresponding organic loading rates (OLR) were 3.76, 4.7, 6.27, 9.4 and 18.8 kg COD/m-3d-1 respectively. 2.3 Analytical procedures Chemical oxygen demand (COD), Total Khjdhal Nitrogen (TKN), Total Phosphorus (TP), Total Solids (TS), Volatile Solids (VS), Suspended Solids (SS), Volatile Suspended Solids (VSS), Alkalinity, Volatile Fatty Acids (VFA) and pH were analyzed. All analytical procedures were performed in accordance with standard methods for examination of water and wastewater APHA [6]. Gas production were measured daily by using water displacement method [7]. The methane content was measured using Gas Chromatograph (GC-8A Shimadzu). 3. Kinetic model 3.1 Monod kinetic model In a UASB reactor without biomass recycle, the rate of change in biomass and substrate concentration in the system can be expressed as equation 1, 2 [8-10]. ௗ௑ ொ௑ ொ௑ = ೚- ೐ + ߤܺ - ‫ܭ‬ௗ ܺ (1) ௗ௧ ିௗௌ

௏ ொௌ

௏ ொௌ

ఓ௑

= ೚- + (2) ௏ ௏ ௗ௧ ௒ The ratio of the total biomass in the reactor to the biomass wasted in a given time period represents the average time that microorganism spend in the reactor. This parameter is called mean cell-residence time (Tc) and is calculated from equation 3 for UASB reactor. ௏௑ (3) ߠ௖ = ொ௑೐

Assuming that the concentration of biomass in the influent can be ignored, at steady-state݀ܺΤ݀‫ = ݐ‬0, and the HRT (TH) is defined as the volume of the reactor divided by the flow rate of the influent. The relationship between the specific growth rate and the rate limiting substrate concentration can be expressed by the Monod as shown in equation 4

93

Sunwanee Jijai et al. / Energy Procedia 79 (2015) 90 – 97 ఓ೘ೌೣ ௌ

ߤ=

(4)

௄ೞ ାௌ

Equation 1 reduces to = ܺሺߤ - ‫ܭ‬ௗ )

ொ௑೐ ௏

ଵ

ߤ=

ఏ௖ ఓ೘ೌೣ ௌ ௄ೞ ାௌ

+ ‫ܭ‬ௗ ଵ

=

ఏ௖

(5) (6)

+ ‫ܭ‬ௗ

(7)

Under steady-state conditions, the rate of change in substrate concentration (݀ܵΤ݀‫ )ݐ‬is negligible and, by a similar technique to that used for the substrate concentration, equation 2 can be reduced to equation 8 by substituting equation 6 ௌబ ିௌ ௑ ଵ = ቀ ൅ ‫ܭ‬ௗ ቁ (8) ఏಹ

௒

ఏ೎

The kinetic parameters Y and ‫ܭ‬ௗ can be obtained by rearranging equation 8 as shown below: ௌబ ିௌ ଵ ௄ = ൅ ೏ ఏಹ ௑

௒ఏ೎

௒

(9)

By plotting equation 9, the values of Y and ‫ܭ‬ௗ can be calculated from the slope and intercept of the best fit line. The value of ߤ௠௔௫ and ‫ܭ‬ௌ could be determined by plotting equation 10, which was derive by rearranging equation 7. Finally, by arranging equation 7, equation 11 is obtained that is used to predict effluent substrate concentration in the reactor. ఏ಴ ௄ ଵ ଵ = ೄ + (10) ଵାఏ಴ ௄೏ ఓ೘ೌೣ ௌ ఓ೘ೌೣ ௄ೄ ሺଵା௄೏ ఏ಴ ሻ

ܵ=

ఓ೘ೌೣ ఏ಴ ି௄೏ ఏ಴ ିଵ

(11)

3.2 Contois model Similar to Monod model, in Contois model, the relationship between the specific growth rate and the rate limiting substrate concentration can be expressed by the equation 12 as shown below: ఓ ௌ ߤ= ೘ೌೣ (12) ఉ௑ାௌ

Again the kinetic parameters Y and ‫ܭ‬ௗ can be calculated from the slope and intercept of the best-fit line equation 9. The value of ߤ௠௔௫ and ߚ could be determined by plotting equation 14. By substituting equation 12 into equation 9 and rearranging we obtain equation 13 and 14. Finally, we can predict the effluent substrate concentration of reactor using equation 15. ఓ೘ೌೣ ௌ ଵ = + ‫ܭ‬ௗ (13) ఉ௑ାௌ ఏ಴

ఏ௖

=

ఉ

௑

+

ଵ

ଵାఏ಴ ௄೏ ఓ೘ೌೣ ௌ ఓ೘ೌೣ ఉ௑ሺଵା௄೏ ఏ಴ ሻ

ܵ=

ఓ೘ೌೣ ఏ಴ ି௄೏ ఏ಴ ିଵ

(14) (15)

3.3 Grau second-order multicomponent substrate removal model The general equation of a second-order Grau model is illustrated in equation 16 [8-11]. ିௗௌ ௗ௧

ௌ ଶ

= ݇௦ ܺ ቀ ቁ

(16)

Integrating equation 16 and then linearizing it, equation 17 is obtained. ௌ = ߠு + బ

(17)

ௌబ

ௌబ ఏಹ

ௌబ ିௌ ௌబ ఏಹ ௌబ ିௌ

௞ೞ ௑

= ܽ + ܾߠு

ܵ = ܵ଴ ቀͳ െ

(18) ఏಹ

௔ା௕ఏಹ

ቁ

(19)

3.4 Modified Stover-Kicannon model In this model the substrate utilization rate is expressed as a function of the organic loading rate using monomolecular kinetics. A special feature of the Modified Stover/Kicannon model is it uses the total organic loading rate as the major parameter to describe the kinetics of an anaerobic reactor in terms of

94


organic matter removal and methane production [9, 10]. Rate of change in substrate concentration in modified Stover̻Kicannon model is shown in equation 20. ௗௌ ௗ௧ ௗௌ ௗ௧

=

ோ೘ೌೣ ሺொௌబ Τ௏ ሻ

(20)

௄ಳ ାሺொௌబ Τ௏ ሻ

Where ݀‫ ݏ‬Τ݀‫ ݐ‬is defined as ொ = ሺܵ଴ െ ܵሻ ௏

௏

=

ொሺௌబ ିௌሻ

ܵ = ܵ଴-

௄ಳ

ቀ

௏

ோ೘ೌೣ ொௌబ ோ೘ೌೣ ௌ೚

ቁ ൅

ଵ ோ೘ೌೣ

௄ಳ ାொௌబ Τ௏

(21) (22) (23)

4. Results and discussion 4.1 Reactor Performance The UASB reactors were operated at five different hydraulic retention times (HRTs), namely 5, 4, 3, 2 and 1 days. The steady state COD removal efficiency decreased from 91.49 to 43.23 %, 89.36 to 45.13 % and 87.23 to 32.69 % for R1, R2 and R3 respectively, with decreasing hydraulic retention times from 5 days to 1 day. This was because the decreasing of HRT led to shorter residence time, thus lowering the efficiency. It also showed that the sizes of granule had a strong effect on the process performance and substrate removal in the UASB reactors. The UASB reactor R1 used only granules from cassava factory which had biggest average sizes (1.5-1.7 mm) gave the highest substrate removal and biogas production. For the UASB reactors R1 and R2 COD removal efficiency became less than 60% at 1 day HRT with OLR of 18.8 KgCOD/m3d, while for R3 the COD removal efficiency became less than 60% at 2 day HRT with OLR of 9.4 KgCOD/m3d because of the disintegration and wash away of biomass or granules along with the effluent due to high mixing intensities [12, 13]. Moreover, as the HRTs decreased from 5 to 1 days, the pH in all three reactors decreased from 7.2 to 5.8, the alkalinity in all three reactors (R1, R2 and R3) decreased from 3,266.67 to 2,404.17, 3,283.33 to 2,471.53 and 2,858.33 to 2,316.67 mg/lasCaCO3 respectively. In contrary the volatile fatty acid (VFA) increased with decreasing HRTs for R1, R2 and R3 (223.61 to 2,433.17, 215.28 to 2,612.50 and 145.14 to 2,554.17 mg/lasCaCO3 respectively). This indicated that acidogenesis occurred more rapidly than acitogenesis and methanogenesis, thus the later steps became the rate-limiting steps which dictated the performance of UASB reactors. 4.2 Kinetic model All kinetic parameters for all four models obtained by fitting stead- state experimental data to the model equations (equation 9, 10, 13, 14, 17, 18 and 22) are shown in table 2, 3. These results were similar to the study of Abtahi et al, 2013, Isik and Sponza, 2005 and Mullai et al, 2011[8, 10, 12]. Table 2 summarizes the kinetic constants determined from Monod and Contois models in previous studies as compared to this study. It should be noted that the substrate concentration used in this study were many times greater than that of most previous studies and thus substrate inhibition may have a significant effect on the reactor performance. In general, the biomass yield in term of Y (gVSS/gCOD) was an order of magnitude lower in our work than that of most previous works. This can mislead to conclude that there were little new cell generation here since methane production is strongly growth associated. The better explanation is that in our cases the cell density is so high (because of the formed granules) such that the limiting substrate could not be supplied throughout the granule mass, causing partial cell starvation, death and lysis which almost balance with new cell generation. Moreover, it was found that UASB-R1 (biggest size in the range 1.5-1.7 mm.) gave the lowest death rate constant value (Kd), whereas UASB-R3 which had smallest granule size showed highest Kd. It should be stated here that Y, μmax and Kd are apparent values rather than the true ones. That is why we see that these values in our work were much lower than that of other works particularly for UASB-R1 which had highest granules size. The true values should be much higher than these values but, because both cell generation, washout and lysis occurred simultaneously and with similar rates, and we saw only the net effect thus small apparent growth (μmax) and dead rate (Kd). Table 2. Comparison of kinetic constant in Monod and Contois Models

95


Monod Model

Reactor

COD (mg/l)

UASB

3,000

UASB

4,214

0.083 -0.83 0.254.17

Y OLR (gVSS/ 3 (kgCOD/m d) gCOD)

Kd (d-1)

μmax (d-)1

Ks (g/l)

CH4yield (mlCH4/ gCODadded)

Reference

6-34

0.78

0.093

0.213

0.560

104-404

1.0-15.8

0.125

0.0065

0.105

!4

104-404

[10] In this study In this study In this study

[9]

18,800

1-5

3.76-18.8

0.0059

0.0021

0.051

19.89

16.6284.56

18,800

1-5

3.76-18.8

0.0089

0.0036

0.012

2.349

5.63-53.99

18,800

1-5

3.76-18.8

0.0280

0.0168

0.024

0.786

6.46-36.56

Reactor

COD (mg/l)

HRT (d)

Y OLR (gVSS/ (kgCOD/m3d) gCOD)

Kd (d-1)

μmax (d-)1

Ks (g/l)

UASB

704.55

0.044-0.848

0.608

0.164

0.0132

0.0212

-

UASB

4,214

1.0-15.8

0.125

0.0065

0.105

0.465

104-404


UASBR1 UASBR2 UASBR3 Contois Model

HRT (d)

UASBR1 UASBR2 UASBR3

0.914.4 0.254.17

CH4yield (mlCH4/ gCODadded)

18,800

1-5

3.76-18.8

0.0059

0.0021

0.023

2.165

16.6284.56

18,800

1-5

3.76-18.8

0.0089

0.0036

0.015

0.674

5.63-53.99

18,800

1-5

3.76-18.8

0.0280

0.0168

0.025

0.169

6.46-36.56

Reference [8]

Other interesting results were the saturated constant (Ks) in Monod model and kinetic parameter (β) in Contois models. Whereas high β in UASB-R1 reflected strong negative effect of granule size on the substrate accessibility of microbial cells, Ks in Monod model indirectly showed similar trend albeit with different interpretation. Highest Ks values in UASB-R1 was a combined effect of high cell density (thus high substrate demand) and diffusion-limiting in-granule substrate transport, both were a result of the biggest granule size. The apparent high Ks value is again does not explain high half-substrate consumption rate according to Monod model formulation. So there is no surprise that both Monod and Contois models could not represent the experimental results so well since their basis in model formulation do not directly include diffusion-limiting step into consideration. Consequently, Monod and Contois are fundamentally unsuitable for modeling UASB with granules and some diffusion-limited consideration must be included in the formulation. Table 3 summarizes the kinetic constants determined for Grau second-order and Modified StoverKicannon models in previous studies as compared to this studies. The kinetic value (KB, Rmax) in Modified Stover-Kicannon model suggest that the substrate removal rate is strongly affected by the granule size. The kinetic value of this model (KB, Rmax) all three reactor (R1, R2, R3) were lower than studies of Büyükkamaci and Filibeli, 2002 (S0=2,000-15,000 mg/l) [14]. But higher than the study of Abtahi et al. (S0=704.55 mg/l) [8] and Sponza and Uluköy, 2008 [9] (S0=3,000 mg/l). The different of kinetic coefficients value because each studies different of initial substrate (S0) and different of granule/inocula that use in each reactors [15]. Table 3. Comparison of kinetic constant in Grau second-order and Modified Stover-Kicannon Models Grau OLR ks COD HRT CH4yield second (kgCOD/ (d-1) Reactor a b (mg/l) (d) (mlCH4/gCODadded) 3 order m d) 0.083UASB 3,000 6-34 0.0291 0.0113 0.26 104-404 0.83 0.25UASB 4,214 1.0-15.8 0.562 1.095 0.337 104-404 4.17 UASB18,800 1-5 3.76-18.8 0.4 1.014 8.23 16.62-84.56 R1 UASB18,800 1-5 3.76-18.8 0.89 0.943 3.42 5.63-53.99 R2 UASB18,800 1-5 3.76-18.8 2.15 0.969 1.55 6.46-36.56 R3

Reference [9] [10] In this study In this study In this study

96


Modified StoverKicannon

Reactor

COD (mg/l)

UASB

3,000

UASB

4,214

UASBR1 UASBR2 UASBR3

HRT (d) 0.0830.83 0.254.17

OLR (kgCOD/ m3d)

KB (gCOD/l.d)

Rmax gCOD/ l.d

CH4yield (mlCH4/gCODadded)

Reference

6-34

0.034

0.0075

104-404

[9]

1.0-15.8

8.211

7.501

104-404


18,800

1-5

3.76-18.8

48.24

47.62

16.62-84.56

18,800

1-5

3.76-18.8

20.06

21.28

5.63-53.99

18,800

1-5

3.76-18.8

6.11

8.77

6.46-36.56

4.3 Model evaluation The kinetic values can be used to explained performance of UASB reactors and to predict the effluent COD from UASB reactors. From four kinetic models (Monod, Contois, Grau-second order, Modified Stover-Kicannon model), when compared with data from experiments, Grau-second order and Modified Stover-Kicannon models were more suitable for predicting COD effluent as indicated by higher regression coefficients (higher than 0.98) for all the rectors (R1, R2 and R3) Actual

7000 6000 5000 4000 3000 2000 1000 0

Contois model

7000 6000 5000 4000 3000 2000 1000 0

1000

yGrau = 1.003x - 21.93 R² = 0.997

2000

3000

4000

1000

2000

3000

7000

4000

5000

6000

7000

R3 yGrau = 0.971x + 148.4 R² = 0.985

yContois = 1.398x - 1491 R² = 0.916

2000

6000

yGrau = 0.967x + 96.65 R² = 0.981 yModified = 0.964x + 78.43 R² = 0.981

yMonod= 1.652x - 2377 R² = 0.890

0

5000

R2

yContois = 1.189x - 558.0 R² = 0.978

0

Modified

yModified = 0.995x - 40.35 R² = 0.997

yMonod = 1.347x - 1002 R² = 0.961

14000 12000 10000 8000 6000 4000 2000 0

Grau second order

R1

yContois = 2.329x - 1537 R² = 0.996

0 Predicted COD (mg/l)

Monod model

yMonod = 0.956x + 95.40 R² = 0.996

4000

yModified = 0.970x + 141.7 R² = 0.985

6000

8000

10000

12000

14000

Measure COD (mg/l)

Fig. 1.Comparison of the predicted and measure COD value from lab scale UASB rectors 5. Conclusion The results of this study showed that sizes of granules highly affected the reactor performance and biogas production in UASB reactors. The COD removal efficiencies and biogas production decreased with decreased HRTs. The experimental data of UASB reactor treating cassava wastewater were in good agreement with the kinetic models particularly Grau second-order and Modifield Stover-Kicannon models.

Acknowledgements The authors would like to thank Walailak University and Ministry of Science and Technology of Thailand grant for financial support.


References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15]

Paepatung. N, Nopharatana. A, Songkasiri W. (2009). Bio-Methane Potential of Biological Solid Material and Agricultural Wastes. Asian Journal on Energy and Environment, 10(1):19-27. Rajagopal R, Saady NMC, Torrijos M, Thanikal JV, Hung Y-T. (2013). Sustainable Agro-Food Industrial Wastewater Treatment UsingHi gh Rate Anaerobic Process. water; 5(1):292-311. Wongnoi R, Phalakornkule C. (2006). Efficiency Enhancement of Up-flow Anaerobic Sludge Bed (UASB) by a Modified Three-phase Separation. Asian Journal on Energy and Environment, 7(3):378-86. Sekiguchi Y, Kamagata Y, Harada H. (2001). Recent advances in methane fermentation technology. Current Opinion in Biotechnology, 12(3):277-82. Habeeb SA, Latiff AABA, Daud ZB, Ahmad ZB. (2011). A review on granules initiation and development inside UASB Reactor and the main factors affecting granules formation process. International Journal of Energy and Envieonment, 2(2):311-20. APHA A, WEF, editor. (1999). Standard Method for the Examination of Water and Wastewater. 20th ed. Washington D.C.: American Public Health Association. Abdel-Hadi M. (2008). A simple apparatus for biogas quality determination. Misr J Ag Eng, 25(3):1055-66. Abtahi SM AM, Nateghi R, Vosoogh A, Dooranmahalleh MG. (2013). Prediction of effluent COD concentration of UASB reactor using kinetic models of monod, contois, second-order Grau and modified stover-kincannon. Int J Env Health Eng, 1(8):1-8. Sponza DT, Uluköy A. (2008). Kinetic of carbonaceous substrate in an upflow anaerobic sludge sludge blanket (UASB) reactor treating 2,4 dichlorophenol (2,4 DCP). Journal of Environmental Management, 86(1):121-31. Işik M, Sponza DT. (2005). Substrate removal kinetics in an upflow anaerobic sludge blanket reactor decolorising simulated textile wastewater. Process Biochemistry, 40(3–4):1189-98. Bhunia P, Ghangrekar MM. (2008). Analysis, evaluation, and optimization of kinetic parameters for performance appraisal and design of UASB reactors. Bioresource Technology, 99(7):2132-40. Mullai Pandian, Huu-Hao NGO, Pazhaniappan S. (2011). Substrate Removal Kinetics of an Anaerobic Hybrid Reactor Treating Pharmaceutical Wastewater. Journal of Water Sustainability, 1(3):301-12. Lettinga G, van Velsen AFM, Hobma SW, de Zeeuw W, Klapwijk A. (1980). Use of the upflow sludge blanket (USB) reactor concept for biological wastewater treatment, especially for anaerobic treatment. Biotechnology and Bioengineering, 22(4):699-734. Büyükkamaci N, Filibeli A. (2002). Determination of kinetic constants of an anaerobic hybrid reactor. Process Biochemistry, 38(1):73-9. Diamantis V, Aivasidis A. (2010). Kinetic analysis and simulation of UASB anaerobic treatment of a synthetic fruit wastewater. Global NEST Journal, 12(2):175-80.

97