Biogas Production Optimization in the Anaerobic Codigestion Process: A Critical Review on Process Parameters Modeling and Simulation Tools

Table 5

Summary of the application of the AI approach in biogas prediction.

AI type	Objectives	Input parameters	Optimum models	Output	Ref.
ANN	To develop a model to forecast the effects of process parameters	COD, BOD, flow rate, recirculated sludge flow, and feed rate	ANN-MLP	Biogas yield	[96]
ANN	To develop a numerical simulation model to estimate the optimum biogas production	Feed type, volatile solid (VS), pH, OLR, HRT, temperature, and reactor volume		Cumulative biogas yield	[97]
ANFIS and ANN	To estimate the biogas yield	C/N ratio, reactor temperature, and RT	ANFIS	Biogas	[24]
ANN	To simulate and model the performances of the codigestion process	Mix ratios and RT	ANN-Bayesian regularization algorithm	Methane yield	[98]
ANN	To model the relationship among the physicochemical parameters of a blend of codigestion of cattle dung and poultry droppings to estimate biogas production	Mix ratios, pH, TDS, temperature, slurry, BOD, and DO		The volume of biogas yield	[94]
ANN-GA and ACO	An integrated model to predict and evaluate the biogas production rate	TSs, VSs, acid detergent fiber, acid detergent lignin, NH3-N, VFA, HRT, and OLR	ANN with GA and ACO	Biogas production rate	[99]
ANN-GA	Optimization and prediction of the amount of biogas generation from codigestion of selected substrates	pH, C/N ratios, and SRT	ANN optimized with GA	Biogas generation	[92]
ANN-GA	To model and optimize mixing ratios in the codigestion process	Substrate-to-inoculum ratios, mix ratios, temperature, SRT, and feed type	ANN-GA	CH4 production	[100]
RNN	To estimate the biogas production rate	SRT, soluble COD, total VFA, free ammonia, and total ammonia		Biogas production rate	[101]
ANN-PSO	To optimize and forecast the biogas generation from codigestion of cattle manure (CM) and palm oil mill effluent (POME)	Mixture ratio, oxidation by hydrogen peroxide, and ammonium bicarbonate		Biogas yields	[102]

RNN, recurrent neural network; SRT, solid retention time.