Development of a hybrid model for sodium gluconate fermentation byAspergillus niger

Abstract

BACKGROUND A hybrid model for sodium gluconate fermentation by Aspergillus niger was proposed. First, because of the lack of part mechanism knowledge on mycelium growth in the mechanism model of the sodium gluconate fermentation process, a back propagation neural network (BPNN) was utilized to develop a model of mycelium growth rate. Second, a three-layer feed-forward network combined with the Alopex-differential evolution (Alopex-DE) algorithm was employed to develop a model of kinetic parameters given that mechanism model mismatch exists in different fermentation batches. A hybrid model based on these two artificial neural network (ANN) models and the mechanism model was developed for the fermentation process. RESULTS The BPNN and ANN models were capable of developing relationships between a certain input and the mycelium growth rate and kinetic parameters. The reliability of the proposed hybrid model was investigated based on 18 batches of experimental fermentation data. Satisfactory results were obtained. CONCLUSIONS The proposed hybrid model is a solution to the mechanism model problems of missing mechanism knowledge and model mismatch in different batches. The model exhibits better performance than pure mechanism and pure BPNN models. © 2013 Society of Chemical Industry