Information Fusion Technique and Its Application to Modeling for Fermentation Processes

For the modeling of the penicillin fermentation processes, a new algorithm, which is based on multiple model fusion technique to modeling of fermentation processes, is proposed. The algorithm can simultaneously use on-line and offline parameters to modeling. Firstly, multiple model fusion-modeling algorithms are designed, the modeling method of two sub models based on adaptive fuzzy neural networks and fuzzy inference are described. Secondly, for three nonlinear testing functions, GMDH-PTSV, Fourier neural network and multiple model fusion-modeling algorithms are used to modeling respectively. Finally, the algorithms of the multiple model fusion modeling for penicillin fermentation process are given. For real data of fermentation processes, the simulation results show that modeling accuracy of the algorithm is better