In the present study, eleven flexible models were employed to describe the effect of different medium compositions (DMC) on ethanol fermentation in repeated-batch biofilm reactors with carob extract. Residual-sum of square, root-mean-square-error, mean-absolute-error, determination coefficient, bias factor, accuracy factor, F test, and objective function were used to compare the models. Findings indicated that corresponding with the prediction of the experimental data of substrate concentration (S), the best-selected models were the Baranyi model (media A and C), Weibull model (medium B), and re-modified Gompertz model (R-MGM) (medium D). It was also found that in the estimation of the observed biomass concentration (X) data, Baranyi model (medium A), Weibull model (medium B), and Stannard model (media C and D) gave well-directed results according to the model comparison, validation, and fitting results. As related to ethanol concentration (P), the predicted data with the re-modified Richards model (R-MRM) (media A and B), re-modified logistic model (R-MLM) (medium C), and Baranyi model (medium D) were showed good agreement with the experimental p values. To validate the best-selected models, an independent set of the experimental data for each medium was used and it was found that the independent experimental values were highly compatible with the selected models. Consequently, the best-selected models can serve as universal equations to fit satisfactorily the experimental S, X, and P curves. These models can also be used for further improvement of the carob extract–based bioethanol production process.
All Science Journal Classification (ASJC) codes
- Renewable Energy, Sustainability and the Environment