ICAE 2012, Jul 5-8, 2012, Suzhou, China Paper ID: ICAE2012- A10732
MODELING OF THE BIOGAS PRODUCTION PROCESS- A REVIEW
1Eva Thorin, 1Eva Nordlander, 1Johan Lindmark, 1Erik Dahlquist, 1Jinuye Yan, 1,2Rebei Bel Fdhila 1
School of Sustainable Society and Technology, Mälardalen University, P. O. Box 883, SE-721 23 Västerås, Sweden, 2
ABB AB, Corporate Research, SE - 721 78, Västerås, Sweden eva.thorin@mdh.se
+4621101564
ABSTRACT
Production of biogas by digestion of organic wastes and other feedstock is one of the important technical solutions that contribute to the transform of the energy system from being fossil fuel dependent to renewable energy originated. To be fully commercial and competitive, the production of biogas needs to be further developed and optimized based on the technical, economic and environmental aspects. Thus, comprehensive understanding of fluid dynamics and microbial reactions in the digestion process is necessary to accurately and robustly model, predict and control the biogas production.
In this paper possible pathways for modeling the biogas reactor is discussed based on previous work on anaerobic digestion modeling and modeling of the fluid flow in reactors. Important parameters for modeling biogas production, with a focus on processes using waste as feedstock, are considered. Identification of knowledge gaps for the modeling of the biogas process is performed and how to overcome the obstacles is addressed.
Key words: biogas, anaerobic digestion, modeling, review, waste
NOMENCLATURE
Abbreviations
ADM1 anaerobic digestion model No. 1 ANFIS adaptive neuro-fuzzy inference system ANN artificial neural networks
ATFBR anaerobic tapered fluidised bed reactor CFD computational fluid dynamics
COD chemical oxygen demand CSTR continuous stirred tank reactor DM dry matter
FFBP feed forward back propagation HRT hydraulic retention time OLR organic loading rate OM organic matter SCOD soluble COD
TCOD total chemical oxygen demand TOC total organic carbon
TS total solids
UASB up-flow anaerobic sludge blanket VFA volatile fatty acids
VS volatile solids
Symbols
r correlation coefficient [-] R2 determination coefficient [-]
1. INTRODUCTION
Since the microorganisms in the biogas production process are sensitive to changes in their environment controlling and predicting the process is challenging. Weiland [1] mentions process monitoring and control as improvements needed for further development of the biogas production process. Process modelling can provide a better understanding of a process and its optimal working conditions. A model can also be used to control the process and predict its outcome.
In the biogas production process the organic material, mainly built up of carbohydrates, proteins and fats, is degraded in several steps to simpler compounds by microorganisms. The process can be divided into four main steps: hydrolysis, acidogenic phase, acetogenic phase and methanogenic phase. In the first step the non-water soluble carbohydrates, proteins and fats are broken down to simple sugars, fatty acids and amino acids outside the microorganisms. The products from the first step are then taken up by the microorganisms and are further degraded into short-chain organic alcohols, hydrogen and carbon dioxide. In the third step acetate is formed. In the last step methanogenic microorganisms are forming methane mainly from acetate, carbon dioxide and hydrogen gas. [2, 3].
The efficiency of the process is influenced by the availability of the microorganisms to the substrate which is dependent on the mixing in the biogas reactor and also by the type and structure of the feedstock as well as the operation condition.
In this paper previous modelling approaches are reviewed and similarities to modelling of the paper pulp process are discussed.
2. ANAEROBIC DIGESTION MODELLING
The development of mathematical models for describing anaerobic digestion started already in 1960´s. A model developed by Andrews in 1969 is considered to be one of the first [4].The first presented models were based on the concept of a rate-limiting step and the aceticlastic methanogenic process step as the limiting one. Later it was found that this is not always true and the first step, the hydrolysis was instead considered to be limiting [5].
How the substrate is characterized is a central aspect in the biogas production modelling and three models using advanced characterization are Angelidaki et al. model, Siegrist et al. model and Anaerobic Digestion Model No. 1(ADM1) [5]. Angelidaki et al characterize the substrate by dividing it into an organic and inorganic part where the organic part includes carbohydrates, proteins, lipids and their degradation intermediates and the inorganic part includes ammonia, phosphate, carbonate, hydrogen sulfide, anions and cations. The conversion of the substrate to biogas is then described with six kinetic equations defining six conversion steps; (1) hydrolysis, (2) acidogenic glucose-degrading, (3) lypolytic, (4) LCFA-degrading acetogenic, (5) VFA- degrading acetogenic and (6) aceticlastic methanogenic steps [6].A more detailed description of the development of kinetic based models from one rate-limiting step to more complex models can be found in [7, 8, 9].
2.1 THE ADM1 MODEL
ADM1 is the result of a common effort by several researchers to form a common platform for modelling of anaerobic digestion. Both Angelidaki and Siegrist participated in this work [10].
The IWA Task Group for Mathematical Modelling of Anaerobic Digestion Processes was formed in 1998 within the International Water Association (IWA) to create a common platform for anaerobic process modelling and simulation. The model was first presented in 2001 and is called Anaerobic Digestion Model No. 1 (ADM1). Both biochemical processes (involving living organisms) as well as physicochemical processes (not involving living organisms) is included in the model. The biochemical processes includes disintegration, hydrolysis, acidogenesis, acetogenesis and methanogenesis and the physicochemical processes concerns liquid-gas processes (liquid-gas transfer) and liquid-liquid processes (ion association/dissociation). Solid-liquid processes are not included due to difficulties in describing them [10].
The ADM1 has been used and modified by several researchers to model different types of biogas production processes and different substrates for example a two-step digester with a thermophilic pre-treatment step and a mesophilic main treatment step, co-digestion of organic waste with activated sludge, co-digestion of cattle manure and energy crops, dry digestion of the organic fraction of municipal waste [11, 12, 13, 14, 15].
One of the greater difficulties with ADM1 is that detailed characterization of the substrate is needed requiring measurements that are not usually made when investigating a biogas plant or wastewater treatment plant and that might not be possible to do on a regular basis. [16, 17, 18]. There have been some attempts to circumvent this problem. One method is to characterize the substrate by measurements of Chemical Oxygen Demand (COD), Total Organic Carbon (TOC), organic nitrogen (Norg) and alkalinity (Alk). The problem to use COD for characterisation of more solid substrates has also been pointed out [14]. There are also suggested procedures including general transformer models to interface ADM1. [17, 19, 20]
Other weaknesses in ADM1 that has been pointed out are inaccuracies in the stoichiometry, some problems with the solid retention time and that there is a lack of restrictions for the thermodynamic boundaries [17].
2.2 MODELLING BASED ON EMPIRICAL
VALUES
For prediction of process performance during different operation conditions and for optimization of processes models based on correlation of empirical values can be used. Khataee and Kasiri [21] made a review of studies on modelling of biological water and waste water treatment processes with Artificial neural networks (ANN). They conclude that ANN models can predict the behaviour of the processes based on experimental data with high correlation coefficients and that additional information about the mechanisms and kinetics of the biological reactions is not necessary. However even though the review by Khataee and Kasiri includes some studies on anaerobic digestion processes no study has the biogas production as output parameter. Some examples on studies on modelling of biogas production based on empirical data are shown in table 1.
In a previous study [22] we have used statistical based models (ANN) to model the biogas production in a full-scale biogas digester using process data from several years of running the digester. The results of this study showed that the general trend and the average level could be followed but the extreme points in the biogas production could not be followed.
According to Waewsak et al. [ 23 ] there are many examples of control systems developed for anaerobic digestion systems but they are often too complex and expensive. They present a control system based on the integration of a neural network model and a fuzzy logic control system which show good results for a lab-scale anaerobic reactor, built as a hybrid of a sludge bed and fixed film reactor.
2.3 OTHER MODELING APPROACHES
There are some examples on simplified models relating the biogas yield to the loading rate of the process or mass balance based. One example is a first order equation used by Yusuf and Ify [24] for co-digestion of cow dung, paper
Table 1 Models for describing biogas production processes based on fitting of empirical values
Model Feedstock/ Process Input Output Validatio
n
Uncertainty Ref. ANN, multilayer,
one and two hidden layers
sludge/ full-scale waste water treatment plant, digester volume 10800
m3
temp., pH, flow rate, VFA, alkalinity, DM,
OM 196 data points effluent VS, CH4 prod. 49 data points r=0.89 for VS r=0.71 for CH4 [25 ANN, multilayer, two hidden layers
organic waste/ full-scale plant (60 tons/day), wet
process temp, TS, VS, pH 177 data points CH4 prod. 50 data points R2=0.87 [26] Fuzzy-logic, MIMO (multiple inputs and multiple outputs) + non-linear regression analysis molasses wastewater/ pilot plant, 0.090 m3, UASB
OLR, TCOD removal rate, alkalinity, pH 134 data points biogas prod., CH4 prod. 40 data points R2=0.98, r=0.97 for both biogas and CH4 R2=0.87, r=0.91 for biogas R2=0.89, r=0.92 forCH4 [27] ANN, multilayer, one hidden layer
molasses/ lab. -scale plant, 0.0075 m3, UASB, thermophilic
OLR, temp. Influent alkalinity + pH, effluent VFA + alkalinity +pH, 60 data points.
biogas prod. 60 data points r= 0.681 r=0.927 for 5-days moving average [28] Multiple linear regression
potato processing waste water/ full-scale plant,
600m3 UASB
flow rate, temp. pH, VFA, alkalinity, influent TCOD + SCOD + temp. + pH, effluent TCOD +
SCOD 2-year historical data
biogas prod. ? χ2 test=
0.28-3.9
[29]
ANN, multilayer, one hidden layer
cattle dung + acetate/lab- scale plant
Influent dung dilution rate or influent acetate
conc., 500 data points and 100 data points
biogas prod. ? MSE = 0.0053-0.0229 [30] Non-linear regression, Levenberg– Marquardt method poultry manure wastewater/pilot scale, 0.0157 m3, UASB HRT, influent COD, 9 data points biogas prod., COD conc. 9 data points (same as model data) R2=0.9954 R2=0.9416 [31]
ANFIS primary sedimentation sludge/ full-scale waste water treatment plant, digester volume 6750 m3
pH, influent VS, flow rate, temp., effluent VS (for biogas prod. model),
132 data points effluent VS biogas prod. 32 data points R2=0.80 R2=0.90 [32]
ANN, one hidden layer
simulated sago industry waste water/lab-scale,
ATFBR, 0.0078 m3
OLR, pH, org. load. removal rate biogas prod. 30 % of data R2=0.9999 R2=0.9997 [33] [34] FFBP neural network
primary and surplus sludge from waste water
treatment/lab-scale CSTR, 0.02 m3
OLR at current and 2 previous time points+ pH, VFA, biogas prod., biogas comp. acetic acid
and propionic acid at previous time point, 500
data points biogas prod. biogas comp. 350 data points regr. coeff.=0.90 regr. coeff.=0.80 [35] [36]
waste and water hyacinth in batch reactors. Mass balance based approaches has also been used for calculation of biogas production [17, 37, 38]. The material balance model described by Kleerebezem and van Loosdrecht [17] has been tested by the authors for a full-scale biogas digester [39] and the results show that the model follows the production trend but the level is hard to predict.
3. MODELLING OF FLUID FLOW
3.1 THE IMPORTANCE OF MIXING
To increase the process efficiency of the biogas production process, one important issue is to increase the reaction rate by optimizing the mixing and gas distribution in the digester. A good and even mixing is important for distribution of microorganisms and nutrition, inoculation of fresh feed, homogenizing of the material and for the removal of end products of the metabolism [3]. Therefore how well the total volume of the digester is being used for the reactions is important for the biogas production.
To get a good biogas production there is a general recommendation from the EPA suggesting a power input of 5 to 8 W/m3 of digester volume [40]. There are researchers that argue that too much mixing could be bad for the process and that a reduction in mixing intensity leads to better reactor performance [41, 42]. The microorganisms are sensitive to too intense mixing which can destroy them [3].Bridgeman [43] studied the impact of mixing on the biogas yield from sewage sludge with a TS of 2.5 % in a laboratory reactor mixed with an impeller. They could not see any difference in biogas production without mixing and with mixing up to 100 rpm.
Mixing in a digester can be performed by mechanical mixing of different sorts, hydraulic mixing by using pumps and pneumatically by using the gas itself for mixing the liquid. Mechanical mixing with different kinds of agitators is the most common type of mixing being used in Europe today [3]. Pneumatically forced circulation stands for about 12 % of the used mixing systems in Europe. There are three types of gas-lift mixing being used in digesters, the free and unconfined release of gas from the bottom, confined gas release inside a draft tube and the use of big bubbles to create a piston pumping action [40]. An advantage of the pneumatically forced circulation is that there are no moving parts in the digester that might break or collect debris and get tangled up.
Research done to study the gas-lift mixing of a digester using a draft tube has revealed that a large portion of the total volume can in fact be poorly mixed [44,45] and especially the surfaces around the bottom. Studies show that poorly mixed zones can be as large as 33,6 % of the digester volume [41]. According to [40] there is no significant reduction in the stagnant zones of the digester by increasing the gas flow rate to three times its original level, so increasing the amount of gas injected is not a good alternative. In the stagnant zones the mixing of new material and microorganisms is low to non-existent and to
utilize the digesters total volume and to avoid sedimentation these stagnant zones should be minimized. Deposit of material at the bottom near to the walls of the digester is to be expected according to experiments done to study the gas mixing within a draft tube [41].
3.2 COMPUTATIONAL FLUID DYNAMIC
MODELLING
The digester is basically a black box with few or no possibilities to evaluate the mixing quality. The hydraulic retention time, the dispersion of the feed, stagnant zones and short-circuiting can be analyzed using for example a tracer method. A chemical tracer (e.g. lithium chloride) is mixed with the injected liquid and by analyzing its concentration at the outlet, the mixing can be evaluated. However, the methods are time-consuming and costly.
Computational Fluid Dynamics (CFD) approach has been proposed by many researchers as an alternative to some of the costly measurements performed on- or off-line. In this simulation method the main design and geometrical details and particularities of the biogas reactor can be taken into account. It is also possible to properly represent the operating characteristics as boundary and initial conditions. [46] and [47] have used a CFD model to study the retention time and dispersion of the feed inside a digester. A tracer has been injected with the feed in the simulation domain to analyze the effect that the mixing has on dispersion of nutrient. This work has provided information on how the distribution of nutrition looks like and short-circuiting effects in a pneumatically mixed digester.
The method is based on the numerical solution of mass, momentum, turbulence and energy equations using finite volumes, finite differences or finite element discretization techniques among others. The main modeling possibilities for the multiphase mixture in the bioreactor are:
1) Eulerian/Eulerian where the injected gas is assumed to be a continua similar to the liquid phase. The phases transport equations are coupled via interphase exchange terms that account for all forces, mass and energy between the gas and the liquid.
2) Eulerian/Lagrangian where the injected gas is treated as dispersed bubbles which are individually tracked in the liquid during the simulation.
3) Volume Of Fluid (VOF) which is an interface tracking method where the interface between gas and liquid is reconstructed and the bubbles and/or gas pockets are automatically generated under the influence of the fluids and flow characteristics as well as the surface tension, [48] and [49]. This method needs a very refined mesh and can be more costly in terms of CPU time compared to the other methods. Meroney and Colorado [46] successfully used a combination VOF/Lagrangian where the top gas in the reactor is considered continuous having a distinct interface with the liquid and the injected gas is assumed dispersed and tracked with a Lagrangian method. Table 2 shows examples of some studies using CFD for modeling biogas processes.
Table 2 Models of fluid flow in biogas reactors
Model Fluid/
Process
Geometry Validation Uncertainty Ref.
CFD, five different turbulence models, non-Newtonian flow sewage sludge/ laboratory vessel with impeller D=0.16 m, H=0.305 m, mixing speeds, 20–200 rpm.
power use for mixing, measurement of torque applied % difference= 13-221 [43] CFD, large eddy simulation, non-Newtonian flow Manure slurry/impeller D=0.152 m, H=0.26 m, 25 angle bottom, 500 rpm computer automated radioactive particle, tracking and computed
tomography
good agreement [50]
CFD, single-phase, laminar flow,
non-Newtonian flow
sludge /full-scale plant, draft tube mechanical mixer,
1100 m3
D=1.1 m, H=15.5 m, cone in top and bottom, mixing rate
8-24 day-1
injected tracer r=0.95 [47]
CFD, standard k–3 turbulence model
sludge/ model + full-scale tanks, single +
multiple draft impeller tube mixers,
1081-8823 m3 D=13.7-33.5 m, H=7.3-10.1 m, cone in bottom, power to volume ratio= 4.1-6.9 W/m3
tracer, HRT good agreement [46]
CFD, three-dimensional steady-state, k–3 turbulence
model (liquid phase)
water/gas lift mixing D=0.2032 m, H=0.295 m, flat and 60 angle bottom, gas velocity 0.024-0.072 cm/s computer automated radioactive, particle tracking, with sludge
good agreement for flow pattern, loc. of dead zones, trends of vel. prof., not god match for, vel.
[51]
To study the mixing dynamics the Växtkraft biogas reactor (The biogas plant of the city of Västerås in Sweden) [48] was simulated for five different flow rates of recirculated gas (air) using a transient, turbulent two-phase flow model. The results were used to visualize and analyze the local process changes as shown in figures 1 and 2. CFD is a promising tool, method and approach where the design geometrical details, the liquid flow, the gas flow and their strong interaction can be implemented in a single package. Integrating the bio-reaction with keeping a reasonable simulation cost is at present the challenge for the scientific community. However, CFD uses a large number of empirical or semi-empirical models where unknown or poorly estimated constants exist. These additional parameters have to be identified by other means. In [52] the local measurements in an air/water experimental facility have been used for validation of a CFD simulation case appropriate for bioreactors or digesters.
4. COMBINATION
OF
MICROBIOLOGICAL
REACTION MODELS AND FLUID FLOW
MODELS
A complete model of the process in a biogas production reactor would include a detailed description of both the microbiological processes degrading the organic material as well as description of the physical transport of material and heat in the reactor.
For solid wastes, such as municipal solid waste, the dry digestion process is an alternative to the wet process. Some
0 0,05 0,1 0,15 0,2 0,25 0 0,2 0,4 0,6 0,8 1 1,2
Gas Flow rate (%)
R e a c to r M e a n Va lu e Liquid Speed
Liquid Turbulent kinetic Energy
Figure 1. The figure shows the mean speed and turbulent kinetic energy of the liquid at different gas flow rates [48].
Figure 2. The figure shows the gas distribution at different flow rates showing the activity at the surface for decreasing gas flow rate from left to right [48]. (The coloured entities are gas bubbles and pockets.)
studies on combining models of the liquid flow through the solid substrate matrix and the biological reactions taking part in the material in the dry digestion process can be found. Nopharantana et al. [ 53 , 54 ] created a model describing the operation of a sequential leach bed for digestion of the organic fraction of municipal waste. The flow is modeled as flows through two interacting domains, one with macro-pores and one with intra-particle micro-pores. The biological reactions are described with a dynamic mass balance model represented by five rate equations. The substrate is represented by two components, an insoluble part and a soluble part.
A mathematical model was proposed by Muha et al. [55]. The model couples microbiological reactions with transport of process liquid and with the variation of population of the microorganisms living on the plastic tower packing of the reactor. Keshtarkar et al [56] combined kinetic models with a two-region liquid mixing model for describing biogas production from cattle manure in a continuously stirred tank reactor. Vavilin et al. [57] also studied the continuously stirred tank reactor and the combination of kinetic models and 2D and 3D transport equations.
CFD modelling has been combined with mass balance based models for plug-flow reactors using manure as feedstock [58]
Another area where integration of fluid flow modelling and kinetic models describing degradation/conversion of material is done is modelling of pulp digesters. In the next chapter some experiences from this area is described.
5. MODELLING OF PULP DIGESTERS
There are several factors that complicate modelling and simulation of the pulp and paper process such as raw material characteristics and that it is hard to quickly and automatically characterise it, time delays in the process, high degree of interaction between production process steps, and that the quality of the end product is dependent on many process parameters in a complex and non-linear way [59].
Blanco et al. [59] give an overview of the use of modelling and simulation in the pulp and paper industry and they also summarise the work within this area made by the research group represented by the authors of this paper. They mention that the advances during later years for simulation of the pulping process have concerned handling multicomponent balances with complex chemical reactions but that still no model can fully describe the processes and that improvements of steady –state modelling is still crucial for achieving better design of pulp and paper processes. They point out the dependency of the quality of the modelling result on the available data. Some data of interest for the modelling is not measured and the expertise needed for analysing the available data can also be lacking. Among the approaches mentioned for modelling are classical statistical methods as well as more advanced methods such as multivariate data analysis and artificial neural networks, physic-chemical deterministic models and
combinations of statistical methods and physic-chemical based methods. Concerning artificial neural networks feed-forward ANNs with back propagation algorithms is said to be the most used in pulp and paper industry.
About 20 simulation software programs used for pulp and paper processes have been found [59] where about half of them are specialised for the pulp and paper process. Further they divide the modelling into two groups: system or process level modelling usually based on mass and energy balance calculations and unit operation level models based on more detailed calculations. Both non-dynamic modelling such as equilibrium thermodynamics and steady-state process simulation and dynamic modelling including reaction kinetics, dynamic process simulation and CFD modelling are used. The latter are used for process analysis and troubleshooting. The level of dynamics in the models used range from tank dynamics, 1D pressure flow dynamics to 3D pressure –flow dynamics. Quality is the most important parameter both for design and operation and models to connecting the properties of the product to the process conditions is still under development and it is mentioned that the complex chemistry complicates the possibility for correlation. For process control two approaches are mentioned: using models based on calibration with real process data and to use physical models tuned with process data.
In previous work done by the research group, of which the authors to this paper are one part, a physical model of a continuous pulp digester has been built. The digester has been divided into five sections which are then divided into five vertical volume elements (2D model). Each volume element is modeled with mass and energy balances based on the inflow, outflow and chemical reactions. The rates of the chemical reactions are correlated to the physical dimension of the wood chips, concentrations of the reaction reactants and products, and temperature. The digestion of the woodchips is modeled as a two-step process with first splitting the lignin from the fibers by hydrogen sulfide and then extracting the lignin in an alkaline environment. The reactivity and the properties of the wood chips are represented by an empirical constant and the temperature dependency of the reaction rate is given by the Arrhenius expression. The input to the model is inflow of woodchips and chemicals, size distribution, and temperature and the output is the amount of lignin that is dissolved which is connected to the quality parameter called Kappa number. The empirical constant is adjusted with the help of measurements of the Kappa number of the fibers leaving the digester. To use the model for diagnostics of the process performance also pressure- flow calculations are included. [59]
A conclusion of Blanco et al. is that to get more robust models for the pulp and paper industry a move from statistical to more physical models is needed.
The main similarity between the biogas digester and pulp digester is that an organic material is degraded in a batch or continues process. The biogas production process is usually small scale compared to the pulp process. Usually a lot of
data is measured on-line in a pulp process while that is not as common in biogas processes. Also automatic control systems are not so well developed for the biogas process. [60]
6. RESULTS AND DISCUSSION –SUGGESTED
FURTHER APPROACH
The development of mathematical models for describing anaerobic digestion has been going on during a long time. The most well-known model including biochemical processes and physicochemical processes is the ADM1 model. However, for practical applications the model requires a too detailed substrate characterisation. Another modelling approach is to use statistical models. The drawback with those models is the difficulty to generalise to several process designs and situations.
There are several similarities between the pulp process and the biogas production process. One of them is the lack in data to be used in detailed modelling of the process.
While in the pulp process modelling the focus is often on product quality it is mainly on product quantity in the biogas process modelling. Still a similarity is the relation of the product quality to complex chemical reactions in the pulp process and the relation of the product quantity to complex microbiological reactions in the biogas process.
In the pulp process modelling a need towards more physical models has been identified while for the anaerobic process the available detailed models have been pointed out to be too complex for practical use. Therefore a combination of empirical and physical/biological models seems to be a possible approach.
7. CONCLUSIONS
A key issue in modelling of biogas production processes is the substrate characterisation. The most well-known model describing the digestion process, ADM1, requires a complex substrate characterisation. Many models based on the ADM1 can be found. Further development of the combination of fluid flow and the degradation reaction models could be a possible way to better describe what is going on in a full scale biogas plant since the contact between the microorganism and substrate and the local environment is important for the biogas production potential of the plant.
There is a need for development of models that can be practical used in full-scale biogas plants for optimization and control.
It would be interesting to also evaluate and further develop the approach of combined empirical and physical/biological models suggested for example for modelling of the pulp process.
REFERENCES
[ 1 ] Weiland P. Biogas production: current state and perspectives, Applied Microbiology and Biotechnology 2010;85: 849–860.
[2] Gerardi H. The Microbiology of Anaerobic Digesters, John Wiley & Sons Inc., Hoboken, New Jersey, 2003
[3] Deublein D, Steinhauser A. Biogas from waste and renewable resources - an introduction, Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA, 2008
[ 4 ] Lübken M, Gehring T, Wichern M. Microbiological fermentation of lignocellulosic biomass: current state and prospects of mathematical modeling, Applied Microbiology and Biotechnology 2010;85:1643-1652
[5] Tomei MC, Braguglia CM, Cento G, Mininni G. Modeling of Anaerobic Digestion of Sludge, Critical Reviews in Environmental Science and Technology 2009;39:1003-1051. [6] Angelidaki I, Ellegaard L, Ahring BK. A comprehensive model of anaerobic bioconversion of complex substrates to biogas, Biotechnology and Bioengineering 1999;63:363-372. [7] Lyberatos G, Skiadas IV., Modelling of anaerobic
digestion - a review, Global Nest: the Int. J. 1999;1,2: 63-76 [8] Donoso-Bravo A, Mailier J, Martin C, Rodrı´guez J, Aceves-Lara CA, Wouwer AV. Model selection, identification and validation in anaerobic digestion: A review, Water Research 2011;45:5347-5364
[ 9 ] Lübken M, Gehring T, Wichern M. Microbiological fermentation of lignocellulosic biomass: Current state and prospects of mathematical modeling, Appl Microbiol Biotechnol 2010;85:1643–1652
[10] Batstone DJ, KellerJ, Angelidaki I, Kalyuzhnyi SV, Rozzi A, Sanders WTM, Siegrist H, Vavilin VA. Anaerobic Digestion Model No. 1 (ADM1), IWA Scientific and Technical Report Series, International Water Association, Cornwall, 2002 [ 11 ] Blumensaat F, Keller J. Modelling of two-stage anaerobic digestion using the IWA Anaerobic Digestion Model No. 1 (ADM1), Water Research 2002;39:171-183. [12] Derbal K, Bencheikh-Iehocine M, Cecchi F, Meniai AH, Pavan P. 2009. Application of the IWA ADM1 model to simulate anaerobic co-digestion of organic waste with waste activated sludge in mesophilic condition, Bioresource Technology 2009;100:1539-1543.
[13] Lübken M, Wichern M, Schlattmann M, Gronauer A, Horn H Modelling the energy balance of an anaerobic digester fed with cattle manure and renewable energy crops, Water Research 2007;41:4085-4096.
[14] Zhou H, Löffler D, Kranert M. Model-based predictions of anaerobic digestion of agricultural substrates
for biogas production, Bioresource Technology 2011;102:10819–10828
[ 15 ] Bollon J, Lehyaric R, Benbelkacem H, Buffiere P. Development of a kinetic model for anaerobic dry digestion processes: Focus on acetate degradation and moisture content, Biochemical Engineering Journal 2011;56:212– 218
[16] Daels T, Willems B, Vervaeren H, Dejans P, Maes G, Dumoulin A, Van Hulle SWH. Calibration and statistical analysis of a simplified model for the anaerobic digestion of solid waste, Environmental Technology 2009;30:.1575-1584.
[ 17 ] Kleerebezem R, Van Loosdrecht MCM. Waste characterization for implementabon in ADM1, Water Science and Technology 2006; 54: 167-174.
[18] Ramirez I, Mottet A, Carrere H, Deleris S, Vedrenne F, Steyer JP. Modified ADM1 disintegration/hydrolysis structures for modeling batch thermophilic anaerobic digestion of thermally pretreated waste activated sludge. Wat. Res. 2009;43,14: 3479-3492.
[19] Zaher U, Buffiere P, Steyer JP, Chen S. A Procedure to Estimate Proximate Analysis of Mixed Organic Wastes, Water Environment Research 2009;81:407-415
[20] Zaher U, Li RP, Jeppsson U, Steyer JP, Chen SL. GISCOD: General Integrated Solid Waste Co-Digestion model, Water Research 2009; 43:2717-2727
[21] Khataee AR, Kasiri MB. Modeling of Biological Water and Wastewater Treatment Processes Using Artificial Neural Networks, Clean – Soil, Air, Water 2011; 39, 8:742–749 [22] Ericson E, Thorin E, Yan J. Exploring the possibility of using a simple neural network for the prediction of biogas production of a solid waste digester, poster presentation at 12th World Congress on Anaerobic Digestion, Guadalajara, Mexico, October 31st – November 4th , 2010
[23] Waewsak C, Nopharatana A, Chaiprasert P. Neural-fuzzy control system application for monitoring process response and control of anaerobic hybrid reactor in wastewater treatment and biogas production, Journal of Environmental Sciences 2010;22, 12: 1883–1890
[24] Yusuf MOL,. Ify NL. The effect of waste paper on the kinetics of biogas yield from the co-digestion of cow dung and water hyacinth, Biomass and Bioenergy 2011;35:1345-1351
[25] Güclü D, Yılmaz N,, Ozkan-Yucel UG. Application of neural network prediction model to full-scale anaerobic sludge digestion, J Chem Technol Biotechnol 2011;86:691– 698
[26] Abu Qdais H, Bani Hani K, Shatnawi N. Modeling and optimization of biogas production from a waste digester using artificial neural network and genetic algorithm, Resources, Conservation and Recycling 2010;54,6:359-363 [27] Turkdogan-Aydınol FI, Yetilmezsoy K. A fuzzy-logic-based model to predict biogas and methane production rates in a pilot-scale mesophilic UASB reactor treating molasses wastewater, Journal of Hazardous Materials 2010;182: 460–471
[28] Kanat G, Saral A. Estimation of Biogas Production Rate in a Thermophilic UASB Reactor Using Artificial Neural Networks, Environ Model Assess 2009; 14:607–614
[ 29 ] Barampouti EMP, Mai ST, Vlyssides AG. Dynamic modeling of biogas production in an UASB reactor for potato processing wastewater treatment. Chemical Engineering Journal 2005 ;106,1 :53-58
[ 30 ] Simeonov I, Chorukova E.. Anaerobic digestion modelling with artificial neural networks, Compt. rend. Acad. bulg. Sci. 2008; 61, 4:505-512
[31] Yetilmezsoy K, Sakar S. Development of empirical models for performance evaluation of UASB reactors treating poultry manure wastewater. Journal of Hazardous Materials 2008;153: 532–543
under different operational conditions
[ 32 ] Cakmakci, M. Adaptive neuro-fuzzy modelling of anaerobic digestion of primary sedimentation sludge, Bioprocess Biosyst Eng 2007; 30:349–357
[33] Parthiban R, Iyer PVR, Sekaran G. Anaerobic digestion in tapered fluidized bed reactorand modeling usingradial basis function neural network, Biotechnology 2007;6, 4,:534-542
[34] Parthiban R, Iyer PVR, Sekaran G. Anaerobic tapered fluidized bed reactor for starch wastewater treatment and modeling using multilayer perceptron neural network, Journal of Environmental Sciences 2007;19:1416–1423 [35] Holubar P, Zani L, Hager M, Fröschl W, Radak Z, Braun R. Advanced controlling of anaerobic digestion by means of hierarchical neural networks, Water Research 2002; 36:2582– 2588
[36] Holubar P, Zani L, Hager M, Fröschl W, Radak Z, Braun R. Start-up and recovery of a biogas-reactor using a hierarchical neural network-based control tool, J Chem Technol Biotechnol 2003; 78:847–854
[37] Namuli R,. Laflamme CB, Pillay P. A Computer Program for Modeling the Conversion of Organic Waste to Energy, Energies 2011;4:1973-2001
[38]Rapport JL, Zhang R, Jenkins BM, Hartsough BR, Tomich TP. Modeling the performance of the anaerobic phased solids digester system for biogas energy production, Biomass and Bioenergy 2011;35:1263-1272
[39] Ericson E, Lindmark J, Thorin E, Yan J, A simplified model for anaerobic digestion of solid waste using real data from a full-scale biogas plant, Venice 2010 -Third International Symposium on Energy from Biomass and Waste, Venice, Italy, November 8-11, 2010
[40] US EPA Process design Manual for Sludge Treatment and Disposal, EPA 625/1-79-011, Cincinnati, Ohio, 1979 [41] Stroot PG, McMahon KD, Mackie RI, Raskin L.Anaerobic codigestion of municipal solid waste and biosolids under various mixing conditions––I. Digester performance, Water Res. 2001;35;1804–1816
[42] Vavilin VA, Angelidaki I. Anaerobic degradation of solid material:Importance of initiation centers for methanogenesis, mixing intesity, and 2D distributed model, Biotechnol Bioengn, 2005;89:113–122
[43] Bridgeman J. Computational fluid dynamics modelling of sewage sludge mixing in an anaerobic digester, Advances in Engineering Software 2012;44:54–62
[44] Karim K, Varma R, Vesvikar M, Al-Dahhan M. Flow pattern visualization of a simulated digester, Water Research 2004;38,17:3659-3670
[45 ]Karim K, Thoma G J, Al-Dahhan MH. Gas-lift digester configuration effects on mixing effectiveness, Water Research 2007;41,14:3051-3060
[ 46 ] Meroney RN, Colorado PE. CFD simulation of mechanical draft tube mixing in anaerobic digester tanks, Water Research 2009;43,4:1040-1050
[47] Terashima M., Goel R, Komatsu K, Yasui H, Takahashi H, Li Y Y and Noike T, CFD simulation of mixing in anaerobic digesters, Bioresource Technology 2009;100,7: 2228-2233. [48 ]Lindmark J, Bel Fdhila R, Thorin E. On Modeling the Mixing in a Digester for Biogas Production, MATHMOD 09, 6th International Conference on Mathematical Modelling, Vienna, Austria, 2009
49 Lindmark J, Bel Fdhila R, Thorin E. Retention time and nutrient tracking inside a digester for biogas production, 52nd International Conference of Scandinavian Simulation Society SIMS 2011, September 29-30, Västerås, Sweden, 2011
[50] Wu B. Large Eddy Simulation of Mechanical Mixing in Anaerobic Digesters, Biotechnology and Bioengineering 2012;109,3:804-812
[51] Vesvikar MS, Al-Dahhan M. Flow Pattern Visualization in a Mimic Anaerobic Digester Using CFD, Biotechnology and Bioengineering 2005; 89,6:719-732
[52] Bel Fdhila R, Simulation of a Bubble Plume in a Water Vessel With and Without Internal Liquid Recirculation, 52nd International Conference of Scandinavian Simulation Society SIMS 2011, September 29-30,Västerås, Sweden, 2011 [53] Nopharatana A, Pullammanappallil PC, Clarke WP. A dynamic mathematical model for sequential leach bed anaerobic digestion of organic fraction of municipal solid waste, Biochemical Engineering Journal 2003;13:21-33 [ 54 ] Nopharatana A, Pullammanappalli, PC, Clarke WP Kinetics and dynamic modelling of batch anaerobic digestion of municipal solid waste in a stirred reactor, Waste Management 2007;27:595-603.
[55] Muha I, Grillo A, Heisig M, Schönberg M, Linke B, Wittum G. Mathematical modeling of process liquid flow and acetoclastic methanogenesis under mesophilic conditions in a two-phase biogas reactor, Bioresource Technology 2012;106:1–9
[56] Keshtkar A, Meyssami B, Abolhamd G., Ghaforian H., Khalagi Asadi M. Mathematical modeling of non-ideal mixing continuous flow reactors for anaerobic digestion of cattle manure, Bioresource Technology 2003;87:113– 1242003
[ 57 ] Vavilin VA, Lokshina LY, Flotats X, Angelidaki I. Anaerobic Digestion of Solid Material:Multidimensional Modeling of Continuous-FlowReactor With Non-Uniform Influent Concentration Distributions, Biotechnology and Bioengineering 2007;97,2:354-366
[58] Wu B., Bibeau EL, Gebremdhin KG. Three-dimesnional numerical simulation modle of biogas production for anaerobic digesters, Canadian Biosystems Engineering 2009;51:8.1-8.7
[59] Blanco A, Dahlquist E, Kappen J, Manninen J, Negro C, Ritala R. Use of modelling and simulation in the pulp and paper industry, Mathematical and Computer modeling of Dynamical systems 2009;15,5: 409-423
[ 60 ] Wiese J, Haeck M. Instrumentation, control and automation for full-scale manure based biogas systems, Water Science and Technology 2006;54,9: 1-8
![Figure 1. The figure shows the mean speed and turbulent kinetic energy of the liquid at different gas flow rates [48]](https://thumb-eu.123doks.com/thumbv2/5dokorg/4899482.134546/5.892.62.849.113.822/figure-figure-shows-turbulent-kinetic-energy-liquid-different.webp)
