Abstract
Modelling has recently become a key tool to promote the bioethanol industry and to optimise the fermentation process to be easily integrated into the industrial sector. In this context, this study aims at investigating the applicability of two mathematical models (Andrews and Monod) for molasses fermentation. The kinetics parameters for Monod and Andrews were estimated from experimental data using Matlab and OriginLab software. The models were simulated and compared with another set of experimental data that was not used for parameters' estimation. The results of modelling showed that mu(max) = 0.179 1/h and K-s = 11.37 g.L-1 for the Monod model, whereas mu(max) = 0.508 1/h, K-s = 47.53 g.L-1 and K-i = 181.01 g.L-1 for the Andrews model, which are too close to the values reported in previous studies. The validation of both models showed that the Monod model is more suitable for batch fermentation modelling at a low concentration, where the highest R squared was observed at S-0 = 75 g.L-1 with an R squared equal to 0.99956, 0.99954, and 0.99859 for the biomass, substrate, and product concentrations, respectively. In contrast, the Andrews model was more accurate at a high initial substrate concentration and the model data showed a good agreement compared to the experimental data of batch fermentation at S-0 = 225 g.L-1, which was reflected in a high R squared with values 0.99795, 0.99903, and 0.99962 for the biomass, substrate, and product concentrations respectively.