Research and Application of Soft Sensor Modeling Method for Total Sugar Content in the Fermentation Process of Chlortetracycline

摘要

At present, the total sugar content in Chlortetracycline fermentation is still analyzed by artificial sampling, the main reason is that fermentation liquid contains a lot of impurities, and its viscosity is larger. It is not possible to measure online by using ordinary sugar content measuring instrument. In order to achieve the on-line prediction of the total sugar content in the fermentation tank, an integrated method based on local modeling and global modeling is proposed in this paper. The integrated soft sensor model of total sugar content is established based on the Just-In-Time Learning (JITL) prediction model, Recursive Least Squares Support Vector Regression (RLSSVR) prediction model and output recurrent wavelet neural network (ORWNN) prediction model. The experimental results show that the integrated soft sensor model can be used to predict the on-line change of total sugar content during the fermentation process, it has high prediction accuracy and lays the foundation for the optimal control in the process of Chlortetracycline fermentation.