Soft Sensor Modeling of Burkholderia cepacia Lipase Fermentation Process Based on BP Neural Network

Abstract:

The aim of this study was to establish a soft sensor model of the Burkholderia cepacia lipase fermentation process.Procedure and method:for this,the backpropagation neural network was applied to rebuild a soft sensor model for the fermentation process of Burkholderia cepacia lipase,and genetic algorithms was further employed to optimize the initial weights and thresholds of neural networks,so as to speed up the convergence rate and achieve a comprehensive optimum result.Result:The model could imitate precisely the process of bacteria growth,culture medium consumption and lipase production,which suggests that the model possesses good reliability and generalization ability.The backpropagation neural network combined with genetic algorithms is an effective and quick method for simulating and predicating fermentation process of B.cepacia lipase.