You have no items in your shopping cart.
ABSTRACT
This study investigated the effect of surfactants (Tween 80, Tween 20, and Triton X-100) on solid state fermentation of citric acid on a tri-substrate feedstock (yam peels, sweet potato peels and cocoyam peels) using Aspergillus niger. The fermentation process was designed using a BoxBehnken design. Response surface methodology (RSM), and Artificial neural network (ANN) and was used to optimized surfactants effects. The analysis of RSM gave a statistically significant quadratic model (p<0.0001) citric acid optimum yield was predicted as 29.3014g/L at Tween 20, Tween 80, Triton X-100; with optimum values of 0.0 g/l, 0.96 g/l and 1.5 g/l respective concentrations. Analysis of ANN showed that a multilayer full feed forward (MFFF) network with quick propagation (QP) and hyperbolic tangent transfer function (Tanh) produced the best model for predicting citric acid yield. The ANN model predicted an optimum citric acid yield of 35.55359g/l at Tween 20, Tween 80, Trinton X-100 as 0.0g/l, 0.54g/l and 1.38 g/l respective concentrations. RSM and ANN models’ predictive capability was assessed based on their RSquared (R2 ) and root mean square error (RMSE) values. These values were obtained as 0.973249 and 1.034549 for RSM and 0.99568 and 1.034549 for ANN respectively. ANN was also considered to be more efficient because of its optimum level of inducers. The inducers were able to enhance lipase production. At maximum lipase concentration of 35.55359g/l, optimum concentrations were Tween 20, Tween 80, Trinton X-100 as 0.0g/l, 0.54g/l and 1.38 g/l. It was therefore concluded that ANN is better than RSM in the optimization and modeling of the effects of inducers on citric acid production.
