Predictive modeling and sensitivity analysis to estimate the experimental data of inulinase fermentation by...

The present work aimed to model Aspergillus niger inulinase fermentation performed in the medium using sigmoidal functions, validate the selected models using an independent set of the experimental values, and perform a sensitivity analysis of the selected models. Based on the results, the selected models were Stannard and Fitzhugh models for substrate consumption (R²  = 0.9976 and 0.9974, respectively), Huang model for inulinase production (R²  = 0.9967), Weibull model for invertase-type production (R²  = 0.9963), and modified logistic model for invertase-type activity/inulinase activity ratio (R²  = 0.9292) with high R² values (>0.90). Kinetics predicted by particularly selected models mentioned above fit well with the experimental kinetic results. Besides, validation of the selected models with an independent set of the experimental data indicated that they gave satisfying results with high R² values for consumption and production (R²  > 0.90). Sensitivity analysis of the selected models showed that the yielded R² values (R²  ≥ 0.9775) were in good agreement with those obtained from the selected models. Consequently, A. niger inulinase fermentation was successfully modeled and the selected models were successfully validated with an independent set of the observed data. Besides, the sensitivity analysis also verified the reliability of the selected models. Those models can serve as universal equations to describe the A. niger inulinase fermentation.