Computational fluid dynamics study on mixing mode and power consumption in anaerobic mono- and co-digestion

Accepted Manuscript Computational fluid dynamics study on mixing mode and power consumption in anaerobic mono- and co-digestion Yuan Zhang, Guangren Yu, Liang Yu, Muhammad Abdul Hanan Siddhu, Mengjiao Gao, Ahmed A. Abdeltawab, Salem S. Al-Deyab, Xiaochun Chen PII: DOI: Reference:

S0960-8524(15)01653-3 http://dx.doi.org/10.1016/j.biortech.2015.12.023 BITE 15844

To appear in:

Bioresource Technology

Received Date: Revised Date: Accepted Date:

10 September 2015 4 December 2015 10 December 2015

Please cite this article as: Zhang, Y., Yu, G., Yu, L., Siddhu, M.A.H., Gao, M., Abdeltawab, A.A., Al-Deyab, S.S., Chen, X., Computational fluid dynamics study on mixing mode and power consumption in anaerobic mono- and co-digestion, Bioresource Technology (2015), doi: http://dx.doi.org/10.1016/j.biortech.2015.12.023

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Computational fluid dynamics study on mixing mode and power consumption in anaerobic mono- and co-digestion Yuan Zhanga, Guangren Yua, Liang Yub, Muhammad Abdul Hanan Siddhua, Mengjiao Gaoa, Ahmed A. Abdeltawabc, Salem S. Al-Deyabc, Xiaochun Chena,* a

College of Chemical Engineering, Beijing University of Chemical Technology, Beijing 100029, PR

China b

Department of Biological Systems Engineering, Washington State University, Pullman, WA 99164,

USA c

Petrochemicals Research Chair, College of Science, King Saud University, Riyadh 11451, Saudi Arabia

* Corresponding author. Tel: +86 10 64433570; Fax: +86 10 64433570. E-mail address: [email protected].

Abstract Computational fluid dynamics (CFD) was applied to investigate mixing mode and power consumption in anaerobic mono- and co-digestion. Cattle manure (CM) and corn stover (CS) were used as feedstock and stirred tank reactor (STR) was used as digester. Power numbers obtained by the CFD simulation were compared with those from the experimental correlation. Results showed that the standard k-ε model was more appropriate than other turbulence models. A new index, net power production instead of gas production, was proposed to optimize feedstock ratio for anaerobic co-digestion. Results showed that flow field and power consumption were significantly changed in co-digestion of CM and CS compared with those in mono-digestion of either CM or CS. For different mixing modes, the optimum feedstock ratio for co-digestion changed with net power production. The best option of CM/CS ratio for continuous mixing, intermittent mixing I, and intermittent mixing II were 1:1, 1:1 and 1:3, respectively. Keywords: computational fluid dynamics (CFD), anaerobic co-digestion, net power production,

1

cattle manure, corn stover. 1. Introduction Anaerobic digestion (AD) is a biological treatment process for different organic wastes to alleviate the energy and environmental issues of society. Due to depletion of non-renewable energy resources and deterioration of global environment, AD is becoming more attractive and popular in society. Anaerobic digestion can be classified as, 1) mono-digestion (i.e., digestion of single substrate) and 2) co-digestion (i.e., simultaneous digestion of two or more substrates). Anaerobic mono-digestion of agriculture waste is yet a problem due to diverse nature and nutrients imbalance of each substrate for anaerobic micro-organisms. Corn stover (CS) is a lignocellulosic biomass with high organic carbon, whereas cattle manure (CM) contains more nitrogen. Thus anaerobic co-digestion of CM and CS can balance the nutrients for micro-organism (Yue et al., 2013), improve the buffer capacity and enhance the methane content and process stability (Belle et al. 2015; Li et al., 2009; Wu et al., 2010; Mata-Alvarez et al., 2011). Co-digestion has higher biogas production and stable performance, e.g., methane production in co-digestion of CM with radish is 39% higher than that in mono-digestion of CM (Belle et al., 2015); biogas production in co-digestion of CM with CS is 24% higher than that in mono-digestion of CM (Yue et al., 2013). Currently, many studies are focused on the anaerobic co-digestion of animal manure and agro-industrial waste (e.g., CM with CS) (Mata-Alvarez et al., 2014; Yue et al., 2013; Li et al., 2014; Li et al., 2009). China is one of the largest agricultural countries and produces different type of wastes such as crop residuals and animal manure. Almost half of the agricultural waste is abandoned or burnt in the open field, while manure is also not utilized and treated fully. This would cause serious environmental and safety problems such as haze weather, fire disaster, and traffic accidents, etc. 2

Converting lignocellulosic biomass to biomethane through anaerobic digestion is an important way to reuse agricultural and animal waste such as CM and CS. Therefore, AD has received increasing attention for treating CM and CS (Li et al., 2014; Li et al., 2009; Yue et al., 2013). Several biogas plants that use CM and CS as feedstock have been constructed in China. For a co-digestion system, the digestion performance is not only influenced by substrates type, feedstock ratio, organic loading rate, and hydraulic retention time but also mixing mode and mixing conditions (Lehtomäki et al., 2007; Li et al., 2014; Wu et al., 2010; Xie et al., 2011b). Mixing is an important operation for distributing nutrients, supplying nutrients to micro-organisms, reducing the inhibitory effect, and adjusting pH in the anaerobic digester. Higher mixing intensity demands more power input and reduces the digester performance. However, there are considerable potential energy savings from lowering mixing intensity and mixing time. It is possible to reduce the power consumption of a biogas plant by using intermittent mixing modes. Intermittent mixing has been shown to be able to produce the same amount of biogas compared to continuous mixing, while decreasing the maintenance and energy demand of the process (Bridgeman, 2012; Karim et al., 2005; Kowalczyk et al., 2013; Lindmark et al., 2014; Ong et al., 2002). Another issue worthy of concerning in mixing of digester is power consumption, which is consumed in motor driven. The power consumption of mixing can vary from 14% to 54% of the total energy demand in a plant (Kowalczyk et al., 2013). With the aim of generating more power in AD process, it is necessary to balance power consumption for AD and energy production from AD. Therefore, it is of great interest to efficiently perform mixing with minimum power consumption and maximum energy production. Thus a new index, net power production is proposed to optimize AD process in this study. Computational fluid dynamics (CFD) is an efficient tool and widely used to study complex

3

phenomena in many disciplines (chemical process, power plants, aeromechanics of vehicles etc) at very low cost, which are not understandable in relatively expensive experimental studies. CFD has been successfully used to study flow field and mixing in anaerobic digester. Some researchers employed CFD to improve impeller efficiency and enhance AD process (Shen et al., 2013; Wu, 2012a; Yu et al., 2011). The effect of other parameters such as substrate loading (Bridgeman, 2012), shear stress (Hoffmann et al., 2008) and mixing rate (Terashima et al., 2009) etc. was also simulated by CFD models. Wu (2011) evaluated six RANS-based (Reynolds-averaged Navier-Stokes) turbulent models in AD mixing simulation, and recommended the standard k-ω and the realizable k-ε models to predict mechanical agitation of non-Newtonian fluids at six total solid (TS) levels. Furthermore, Wu (2012b) suggested that large eddy simulation (LES) performed better than RANS simulation in predicting turbulent flow in anaerobic digester; however, its industrial application was limited due to high computing costs. In previous studies, major efforts have been made on mono-digestion. To our best information, there is no report on the relationship between appropriate substrates ratio of co-digestion and power consumption/production of STR. Therefore, the main objectives of this study were to: (1) develop an appropriate numerical model of CFD for AD in our research; (2) investigate the flow field and power consumption in mono- and co-digestion of CM and CS in STR; (3) determine the effective feedstock ratio of co-digestion and synergistic effect on power production. 2. Materials and methods 2.1. Substrates and AD process In this study, CM and CS were selected as substrates for AD process. Both CM and CS were collected from Shunyi County of Beijing City, China. The CM was diluted with tap water to 4

approximately 12%TS, and then long manure fiber was separated using a sieve with a 1 cm diameter screen. After that, the CM slurry was stored in a freezer (at -20 oC). When used as feedstock for anaerobic digestion, the CM slurry was further diluted to 5.4%TS. The CS was chopped into 1.0-1.5 cm in length after being air-dried, and then the chopped CS was pretreated by 2% NaOH. After washing the pretreated CS, it was air-dried and ground into a particle size of 20-mesh with a lab-scale mill. The ground CS particles were made into slurry with 5.75%TS for anaerobic digestion. The details for the characteristics of CM and CS were provided in the previous research (Li et al., 2009; Tian et al., 2014). In the CFD simulation, CM and CS slurries were assumed to be incompressible and pseudo-plastic fluids. The power law model was used to describe the slurry rheological properties (Achkari-Begdouri and Goodrich, 1992; Landry et al., 2004; Tian et al., 2014; Wu and Chen, 2008). The rheological parameters and the related equations were provided in Table 1. AD process was implemented in sequencing mode in an unbaffled STR with working volume of 8.0 L. The initial substrate loading concentration was 50 g L-1 and the seeding sludge inoculation was 15 g L-1. The seeding sludge was taken from a manure anaerobic digester in Nanwu, Beijing, China. The STR was operated with a hydraulic retention time (HRT) of 40 days at each loading concentration. Two types of mixing modes were tested: continuous mixing and intermittent mixing. There were also two intermittent mixing modes in this study. The first intermittent mixing mode called INTER I means to agitate for 20 min every 2 hours. The second intermittent mixing mode called INTER II means to agitate for 5 min every 2 hours (Tian et al., 2014; Tian et al., 2015). The CM/CS ratios of 1:0, 1:1, 1:2, 1:3, 1:4, and 0:1 were used (Li et al., 2009). The AD process was operated at mesophilic temperature (35± 2 ℃) and 80 rpm agitation speed.

5

2.2. CFD simulation An unbaffled STR was used in the CFD simulation for mono- and co-digestion of CM and CS. The STR reactor was 0.2 m in diameter and 0.32 m in height (liquid surface height of 0.26 m). A stirring motor was mounted on the top of the STR to drive a vertical shaft with three-layered pitched blade. The diameter, width, thickness, and angle of each pitched blade were 0.17 m, 0.04 m, 0.003 m and 45°, respectively. CFD simulation was performed using FLUENT 14.0 software. During co-digestion a two-phase mixture model including two non-Newtonian fluids was used to predict the flow behavior of anaerobic slurry. The CS slurry was considered as the main phase while the CM slurry was considered as the secondary phase. The steady state three-dimensional simulation was implemented. The interaction between gas phase (biogas) and liquid phase (anaerobic slurry) was assumed to be negligible. In the CFD simulation for anaerobic co-digestion of CM and CS, the CS/CM ratios were set up using “Patch” function of FLUENT software. The unstructured grids were generated by the software ICEM 14.0 and the number of the grids was 742,593 cells. The minimum cell volume was 2.01 × 10-7 cm3 and the maximum cell volume was 1.23 × 10-3 cm3. The criteria of the scaled residuals used to stipulate an equilibrium flow were obtained when it decreased to 1×10-5 for continuity, x-velocity, y-velocity, z-velocity, k and ε. Another critical convergence criterion for each solution was that the agitator torque approached a constant value. 2.3. Experimental correlation and numerical simulation of power number Power number is an important parameter for the assessment of agitator mixing performance and it is greatly affected by the slurry’s rheological properties (Chudacek, 1985). In this study, two methods were used to calculate power number of agitator. One is empirical correlation derived from experimental data. Another is numerical simulation derived from theoretical

6

equations. The power number of agitator from empirical correlation was calculated by the following equations (Nagata, 1975). q

N P-exp

 1000 + 1.2 Re0.66  A (0.35+b / D ) ( sinθ )1.2 = + B H / D) 0.66  ( Re  1000 + 3.2 Re 

A = 14 + ( b / D ) 670 ( d / D − 0.6 ) + 185    2

(2)

1.3 − 4( b / D − 0.5 )2 −1.14 ( d / D )  

B = 10 

2

q = 1.1 + 4b / D − 2.5 ( d / D − 0.5 ) − 7 ( b / D )

(1)

(3) 4

(4)

where NP-exp is power number in experimental correlation, H is height from liquid surface to reactor bottom (m), D is diameter of vessel (m), θ is angle of pitched blade (°), d is diameter of agitator (m), and b is width of one-layered pitched blade. The width of blade of three-layered pitched blade is triple that of one-layered one. The power number of agitator from CFD simulation was calculated by the following equations (Wu and Lionel, 2000):

P = 2π NT NP =

P ρ N 3d 5

(5)

(6)

where NP is power number in CFD simulation, P is power consumption of agitator (W), N is agitator speed (rps), and T is agitator torque (Nm) which is obtained from the CFD simulation results. Error indicators were used to quantitatively evaluate the performance of turbulence models in CFD simulation. The error indicator δp of power number was calculated by following equation:

7

δP =

N P -N P -exp N P -exp

× 100%

(7)

3. Results and discussion

3.1. Model validation Table 2 shows the validation of the different turbulence models by comparing the predicted power number with the experimentally correlated one (Nagata, 1975). Reynolds stress with an appropriate turbulence model is critical to characterize the flow field in a constrained vessel with rotating elements. Five turbulence models were assessed by predicting the power number. In an unbaffled STR, the power number was empirically correlated from the experimental data (Nagata, 1975). Empirical correlation has been used to validate the CFD models (Xie et al., 2011a; Wu, B., 2012b). It is reasonable to use this method to evaluate the performance of the turbulence models in anaerobic slurry flow. The correlated power number NP-exp was 5.14 in the anaerobic mono-digestion of CS with 5.75% TS at 80 rpm. Error indicators of power number NP showed that more accurate result was obtained from the standard k-ε model (S k-ε). The lowest error indicator of power number at 5.8% was observed. Therefore, the S k-ε model was considered to be the best option and recommended for the mixing study in this research. 3.2. Anaerobic mono-digestion of CM or CS Fig. 1 (a) and (b) show the quantitative comparison of the flow patterns between CM and CS slurry. Although the impellers and mixing conditions were the same, significantly different flow fields were formed in two different slurries. In the CS digester, large areas at or near the agitator shaft, reactor wall, and spaces between agitators were observed in stagnant zones (velocity magnitude < 0.1 m/s) due to higher viscosity compared to that of the CM slurry. The flow parameters and results such as average viscosity, wall shear stress in the CS and CM slurries are 8

shown in Table 3. Fig. 2 (a) and (b) show the effect of Reynolds number (Re) on power number in mono-digestion of CS and CM. The power numbers of CM and CS were calculated by the aforementioned two methods. Good agreements were obtained between the correlated and predicted values for power number of agitator. In the CS digester, the power number was inversely proportional to Re when Re < 10. Once Re was greater than 100, the power number tended to be stable with Re because the flow gradually changed into turbulent flow. Similar trend was observed for power number in the CM digester when Re increased from 100 to 1000. However, the flow behavior in the CS digester was significantly different from that in the CM digester. Metzner and Otto (1957) reported the flow behavior of carboxyl methyl cellulose (CMC) solution as non-Newtonian fluid. When Re < 10, the CMC flow was in laminar flow. When 10 < Re < 100, transition flow prevailed the flow regime. When Re > 100, turbulent flow dominated the flow regime. According to the analysis of NP-Re relationship shown in Fig. 2 (a) and (b), the results of CS flow were in consistent with that of CMC flow while the results of CM flow were different from that of CMC flow. In the CM digester, when 100 < Re < 1000, transition flow prevailed the flow regime. When Re > 1000, turbulence flow dominated the flow regime. Fig. 3 shows the effect of Re on power consumption in mono-digestion of CS and CM. With an increase of Re, a sharp increase of power consumption for the CS agitator and a slow increase of power consumption for the CM agitator were observed. When Re was 100, the power consumption for the CM agitator was 0.04 W while the power consumption for the CS agitator was up to 3.2 W. Therefore, the power needed for the CS agitator was higher than that for the CM agitator under the same flow condition. 3.3. Prediction of flow characteristics in anaerobic co-digestion of CM and CS

9

Velocity and shear stress are important results for the assessment of mixing performance. The aforementioned results for mono-digested slurry of CM, co-digested slurry of CM and CS and mono-digested slurry of CS are shown in Table 3. Different trends for velocity and shear stress were observed in the three digestion systems. A decrease trend was observed for average velocity and impeller shear stress (CM > co-digestion > CS) while an increase trend was observed for maximum wall shear stress (CM < co-digestion < CS), suggesting that a better mixing was obtained in the CM digester. Fig. 4 shows viscosity distribution in the CS digester and the co-digester (CS/CM 1:1). Uneven distribution of viscosity was observed both in the CS digester (Zhang et al., 2014) and the co-digester of CM and CS due to shear-thinning rheological property. Fig. 5 shows the velocity contour at the axial and cross sectional areas in the co-digester of CM and CS. A small stagnant zone occurred close to agitator shaft and tank wall in the co-digester of CM and CS due to low velocity. The following trend was observed by comparing the stagnant zone area of the co-digester with that of the mono-digester: CM < co-digestion < CS. The larger stagnant zone area occurred in the mono-digester of CS due to the higher viscosity and uneven mixing. In addition, many investigations had shown that under- or over-intensive mixing would undermine stability of AD process (Vavilin V. A. and, Angelidaki I., 2005). Some researchers reported that high mixing intensity had resulted in floc breakage and low gas production due to high shear stress. Furthermore, uneven mixing with stagnant zones worked as initiation centers for methanogens to protect them during acidification (Lindmark et al., 2014). Therefore, stability of co-digestion would prefer the moderate mixing with certain stagnant zones. 3.4. Prediction of power consumption in anaerobic co-digester of CM and CS

10

Fig. 6 (a) shows the effect of agitator speed on power consumption in the mono- and co-digester of CM and CS. The power consumption of co-digester (CS/CM = 1:1) was lower than that of CS mono-digester under each agitator speed. The power consumption showed significant increases with an increase of agitator speed from 20 rpm to 170 rpm. No significant difference in power consumption was observed when agitator speed was less than 80 rpm. However, the difference in power consumption gradually increased with the increase in agitator speed. Compared with mono-digestion of CS, co-digestion of CM and CS would require less agitator power consumption under the same agitator speed. Fig. 6 (b) shows the effect of feedstock ratios on power consumption at the agitator speed of 80 rpm and 160 rpm. At the agitator speed of 80 rpm, the power consumption of agitator was directly proportional to the CS ratio in the co-digestion slurry. At the agitator speed of 160 rpm, growth curve of power consumption was observed up to the feedstock ratio of CS/CM (3:7) and then it changed to linear response. Table 4 shows the results of power consumption for the different feedstock ratio at the speed of 80 rpm. Compared with mono-digestion of CM, the power consumption for co-digestion increased with increase in the CS ratio. As reported by Li et al. (2009), biogas yield at the feedstock ratios (CM/CS) of 1:0, 1:1, 1:2, 1:3, 1:4, and 0:1 were 0.222, 0.283, 0.304, 0.315, 0.312, and 0.326 L/gTS, respectively. Based on the aforementioned research, potential biogas productions at the feedstock ratios of 1:0, 1:1, 1:2, 1:3, 1:4, and 0:1 were 95.9, 126.2, 137.3, 142.5, 141.6, and 145.4 L in the digester working volume of 8 L, respectively. In China, 1 m3 of biogas generally generates 1.6-2.4 kWh electricity with different power equipments (Tian, 2010). In this study, 2.0 kWh /m3 biogas was used, and the power production of AD at different feedstock ratio (CM/CS) was shown in Table 4. To assess power consumption

11

and power production from AD process, a new index called net power production which is the sum of power production (+) and power consumption (-) was proposed to co-digestion system. Net power production under continuous and intermittent mixing conditions was further compared and discussed as follows. The digesters was operated under continuous mixing conditions for 40 days during co-digestion, and net power production at four feedstock ratios of CM/CS (1:1, 1;2, 1:3, 1:4) had a negative decrease with increase in the CS ratio. The results suggested that larger CS ratio required more external power in a co-digestion system. For co-digestion of CM and CS with INTER I, the net power production at four feedstock ratios CM/CS (1:1, 1;2, 1:3, 1:4) had a positive decrease with increase in the CS ratio. Among four feedstock ratios, the co-digestion with CM/CS ratio of 1:1 produced the maximum net power production (+0.0377 kWh). However, the net power production under the INTER II mixing was different from that under the INTER I mixing. Among four feedstock ratios, the co-digestion with CM/CS ratio of 1:3 produced the maximum net power production (+0.2215 kWh). Therefore, the net power production was an appropriate measure to optimize feedstock ratio of co-digestion under three different mixing conditions. The feedstock ratio of co-digestion not only had significant impact on biogas production but also on power consumption. The current motivation was to optimize feedstock ratio based on biogas production and stability for sustainable AD development. The prediction of CFD simulation demonstrated a significant relation between feedstock ratio and power consumption in an anaerobic co-digestion system. This suggests that the future design for anaerobic co-digestion should not only focus on biogas production but also on power consumption to obtain a positive net power production. 4. Conclusions

12

Through comparing power numbers obtained from CFD simulation with those from experimental correlation, standard k-ε turbulence model was selected to evaluate the flow fields and power consumption in STR. Mixing uniformity of mono- and co-digestion were decreased in the following trend: CM > co-digestion > CS. Net power production was proposed to optimize the feedstock ratio of anaerobic co-digestion, and the index indicated that 1:1, 1:1, and 1:3 were proper for continuous mixing, INTER I, and INTER II, respectively. Therefore, power production of co-digestion could be improved through change of mixing mode and optimization of feedstock ratio. Acknowledgements

The authors are grateful for the financial support from National Natural Science Foundation of China (21176021, 21276020) and Basic Scientific Research Foundation for Chinese Universities (JD1301). We extend our appreciation to the Deanship of Scientific Research at King Saud University for funding the work, through Research Group Project No. RG-1436-026. References

1. Achkari-Begdouri, A., Goodrich, P.R., 1992. Rheological properties of Moroccan dairy cattle manure. Bioresour. Technol. 40, 149-156. 2. Belle, A.J., Lansing, S., Mulbry, W., Weil, R.R., 2015. Anaerobic co-digestion of forage radish and dairy manure in complete mix digesters. Bioresour. Technol. 178, 230-237. 3. Bridgeman, J., 2012. Computational fluid dynamics modelling of sewage sludge mixing in an anaerobic digester. Adv. Eng. Software 44, 54-62. 4. Chudacek, M.W., 1985. Impeller power numbers and impeller flow numbers in profiled bottom tanks. Industrial & Engineering Chemistry Process Design and Development 24, 858 867. 13

5. Hoffmann, R.A., Garcia, M.L., Veskivar, M., Karim, K., Al-Dahhan, M.H., Angenent, L.T., 2008. Effect of shear on performance and microbial ecology of continuously stirred anaerobic digesters treating animal manure. Biotechnol. Bioeng. 100, 38-48. 6. Karim, K., Hoffmann, R., Thomas Klasson, K., Al-Dahhan, M.H., 2005. Anaerobic digestion of animal waste: Effect of mode of mixing. Water Res. 39, 3597-3606. 7. Kowalczyk, A., Harnisch, E., Schwede, S., Gerber, M., Span, R., 2013. Different mixing modes for biogas plants using energy crops. Appl. Energy 112, 465-472. 8. Landry, H., Laguë, C., Roberge, M., 2004. Physical and rheological properties of manure products. Appl. Eng. Agric. 20, 277-288. 9. Lehtomäki, A., Huttunen, S., Rintala, J.A., 2007. Laboratory investigations on co-digestion of energy crops and crop residues with cow manure for methane production: Effect of crop to manure ratio. Resources, Conserv. Recycl. 51, 591–609. 10. Li, J., Wei, L., Duan, Q., Hu, G., Zhang, G., 2014. Semi-continuous anaerobic co-digestion of dairy manure with three crop residues for biogas production. Bioresour. Technol. 156, 307–313. 11. Li, X., Li, L., Zheng, M., Fu, G., Lar, J.S., 2009. Anaerobic co-digestion of cattle manure with corn stover pretreated by sodium hydroxide for efficient biogas production. Energy & Fuels, 23, 4635-4639. 12. Lindmark, J., Thorin, E., Fdhila, R.B., Dahlquist, E., 2014. Effects of mixing on the result of anaerobic digestion: Review. Renewable and Sustainable Energy Rev. 40, 1030–1047. 13. Mata-Alvarez, J., Dosta, J., Macé, S., Astals, S., 2011. Codigestion of solid wastes: A review of its uses and perspectives including modeling. Critical Rev. Biotechnol. 31, 99-111. 14. Metzner, A.B., Otto, R.E., 1957. Agitation of non-Newtonian fluids. AIChE J. 3, 3-10.

14

15. Nagata, S., 1975. Mixing: principles and applications. Kodansha, New York. 16. Ong, H.K., Greenfield, P.F., Pullammanappallil, P.C., 2002. Effect of mixing on biomethanation of cattle-manure slurry. Environ. Technol. 23, 1081-90. 17. Shen, F., Tian, L., Yuan, H., Pang, Y., Chen, S., Zou, D., Zhu, B., Liu, Y., Li, X., 2013. Improving the mixing performances of rice straw anaerobic digestion for higher biogas production by computational fluid dynamics (CFD) simulation. Appl. Biochem. Biotechnol. 171, 626-642. 18. Terashima, M., Goel, R., Komatsu, K., Yasui, H., Takahashi, H., Li, Y., Noike, T., 2009. CFD simulation of mixing in anaerobic digesters. Bioresour. Technol. 100, 2228-2233. 19. Tian, L., Shen, F., Yuan, H., Zou, D., Liu, Y., Zhu, B., Li, X., 2014. Reducing agitation energy-consumption by improving rheological properties of corn stover substrate in anaerobic digestion. Bioresour. Technol. 168, 86-91. 20. Tian, L., Zou, D., Yuan, H., Wang, L., Zhang, X., Li, X., 2015. Identifying proper agitation interval to prevent floating layers formation of corn stover and improve biogas production in anaerobic digestion. Bioresour. Technol. 186, 1-7. 21. Tian Yishui, 2010. Biomass power generation, Chemical Industry Press, Beijing (In Chinese). 22. Vavilin V. A., Angelidaki I, 2005. Anaerobic degradation of solid material: importance of initiation centers for methanogenesis, mixing intensity, and 2D distributed model. Biotechnol. Bioeng. 89, 113-122. 23. Wu, B., 2011. CFD investigation of turbulence models for mechanical agitation of non-Newtonian fluids in anaerobic digesters. Water Res. 45, 2082-2094. 24. Wu, B., 2012a. CFD simulation of mixing for high-solids anaerobic digestion. Biotechnol.

15

Bioeng. 109, 2116-2126. 25. Wu, B., 2012b. Large eddy simulation of mechanical mixing in anaerobic digesters. Biotechnol. Bioeng. 109, 804-812. 26. Wu, B., Chen, S., 2008. CFD simulation of non-Newtonian fluid flow in anaerobic digesters. Biotechnol. Bioeng. 99, 700-711. 27. Wu, J., Lionel, P., 2000. Performance analysis of axial-flow mixing agitators. AIChE J. 46, 489 - 498. 28. Wu, X., Yao, W., Zhu, J., Miller, C., 2010. Biogas and CH4 productivity by co-digesting swine manure with three crop residues as an external carbon source. Bioresour. Technol. 101, 4042–4047. 29. Xie, M.H., Zhou, G.Z., Xia, J.Y., Zou, C., Yu, P.Q., Zhang, S.L. 2011a. Comparison of power number for paddle-type impellers by three methods. Journal of Chemical Engineering of Japan, 44(11), 840-844. 30. Xie, S., Lawlor, P.G., Frost, J.P., Hu, Z., Zhan, X., 2011b. Effect of pig manure to grass silage ratio on methane production in batch anaerobic co-digestion of concentrated pig manure and grass silage. Bioresour. Technol. 102, 5728-5733. 31. Yu, L., Ma, J., Chen, S., 2011. Numerical simulation of mechanical mixing in high solid anaerobic digester. Bioresour. Technol. 102, 1012-1018. 32. Yue, Z., Chen, R., Yang, F., MacLellan, J., Marsh, T., Liu, Y., Liao, W., 2013. Effects of dairy manure and corn stover co-digestion on anaerobic microbes and corresponding digestion performance. Bioresour. Technol. 128, 65-71. 33. Zhang, L., Zhang, J., Li, C., Bao, J., 2014. Rheological characterization and CFD modeling of corn stover-water mixing system at high solids loading for dilute acid pretreatment. Biochem.

16

Eng. J. 90, 324-332.

17

Table 1 Rheological parameters and equations for anaerobic slurries

Property TS%（w/w） K（Pa sn） n γ（s-1） η（Pa s） ρ（kg/m3）

CM 5.40 0.19 0.56 50-702 0.01-0.03 1000.78

CS 5.75 4.30 0.29 0.21-40.73 0.31-13.02 1003.00

Reference (Achkari-Begdouri and Goodrich, 1992; Tian et al., 2014)

τij (Pa)

 ∂u ∂u  τ ij =η  i + j   ∂x ∂x  i   j

η (Pa s)

η = K γ n −1

Re

n ρ N 2−n d 2   n   (Metzner and Otto, 1957) Re = 8    K   6n + 2  

Note: operating temperature 35℃. TS, total solids. K is the consistency coefficient, n is the power-law index, γ is the shear rate (s-1), η is the viscosity, τij is the shear stress, Re is Reynolds number, N is agitator speed (rps), and d is diameter of pitched blade (m).

18

Table 2 Comparison of different turbulence models with experimental correlation (NP-exp = 5.14 )

Models NP

S k-ε 5.44

R k-ε 5.56

RNG k-ε 5.45

T k-ω 5.59

RSM 5.63

5.8 8.2 6.0 8.7 9.5 δP /% Note: mono-digestion of CS at 80 rpm. S k-ε ( standard k-ε model ), R k-ε ( realizable k-ε model ), RNG k-ε ( renormalised Group k-ε model ), T k-ω ( transition k–kl-ω model ), and RSM ( Reynolds stress model ).

19

Table 3 Flow parameters and results in mono-digestion and co-digestion reactors

Maximum Parameters velocity (m/s) CM 0.83 Co-digestion 0.74 CS 0.73

Average velocity (m/s) 0.24 0.18 0.16

Average viscosity (Pa s) 0.03 0.52 1.05

20

Maximum wall shear stress (Pa) 31.26 215.17 350.10

Impeller shear stress (Pa) 0.31 0.21 0.09

Table 4 Power consumption for different feedstock ratio at the agitator speed of 80 rpm CM/CS 1:0 1:1 1:2 1:3 1:4 0:1 Power/10-3 kW 0.793 1.342 1.508 1.587 1.629 1.837 Potential biogas production /L 95.9 126.2 137.3 142.5 141.6 145.4 Equivalent power production / kWh +0.1918 +0.2524 +0.2746 +0.2850 +0.2833 +0.2907 Power consumption -0.7611 -1.2884 -1.4483 -1.5246 -1.6290 -1.7642 / kWh Continuous mixing Net power -0.5692 -1.0356 -1.1734 -1.2390 -1.3457 -1.4731 production / kWh Power consumption -0.1269 -0.2147 -0.2413 -0.2539 -0.2606 -0.2939 Intermittent / kWh mixing I Net power (INTER I) +0.0649 +0.0377 +0.0333 +0.0311 +0.0227 +0.0032 production / kWh Intermittent Power consumption -0.0317 -0.0537 -0.0603 -0.0635 -0.0652 -0.0735 mixing II / kWh (INTER II) Net power +0.1601 +0.1987 +0.2143 +0.2215 +0.2181 +0.2174 production / kWh Note: Net power production = Power consumption + Power production. Continuous mixing lasted for 40 days (24 h/day). Intermittent mixing was controlled as 12 times per day and lasted for 20 min (INTER I) or 5 min (INTER II) for each stirring. “+” indicated power output, “-” indicated power input.

21

(a)

(b) Fig. 1. Velocity contour of CM digester (a) and CS digester (b) at 80 rpm.

22

(a)

(b) Fig. 2. Effect of Re on power number in mono-digestion of CS (a) and CM (b).

23

Fig. 3. Effect of Re on power consumption in mono-digestion of CS and CM.

24

(a) CS slurry

(b) Co-digestion slurry

Fig. 4. Viscosity comparison of CS slurry and co-digestion slurry.

25

Fig. 5. Velocity contour of co-digester of CM and CS.

26

(a)

(b) Fig. 6. Effect of agitator speed (a) and CS/CM ratio (b) on power consumption.

27

HIGHLIGHTS


CFD models were developed for anaerobic mono-digestion and co-digestion.




Continuous and intermittent mixing was compared in anaerobic co-digestion.




New index of net power production was proposed to optimize feedstock ratio.




Optimum feedstock ratios were determined for different mixing modes.

28