Modeling dark fermentation for biohydrogen production: ADM1-based model vs. Gompertz model

Biohydrogen production by dark fermentation in batch reactors was modeled using the Gompertz equation and a model based on Anaerobic Digestion Model (ADM1). The ADM1 framework, which has been well accepted for modeling methane production by anaerobic digestion, was modified in this study for modeling hydrogen production. Experimental hydrogen production data from eight reactor configurations varying in pressure conditions, temperature, type and concentration of substrate, inocula source, and stirring conditions were used to evaluate the predictive abilities of the two modeling approaches. Although the quality of fit between the measured and fitted hydrogen evolution by the Gompertz equation was high in all the eight reactor configurations with r2 ∼0.98, each configuration required a different set of model parameters, negating its utility as a general approach to predict hydrogen evolution. On the other hand, the ADM1-based model (ADM1BM) with predefined parameters was able to predict COD, cumulative hydrogen production, as well as volatile fatty acids production, albeit at a slightly lower quality of fit. Agreement between the experimental temporal hydrogen evolution data and the ADM1BM predictions was statistically significant with r2 > 0.91 and p-value <1E-04. Sensitivity analysis of the validated model revealed that hydrogen production was sensitive to only six parameters in the ADM1BM.