Background: How to accurately predict the occurrence of contamination in the fermentation process of Chlortetracycline? How to prompt field operators to take effective measures in time? This is a difficult problem that the fermentation process of Chlortetracycline has not been solved well.
Objective: The aim of this paper is to effectively predict whether the fermentation process of Chlortetracycline is contaminated or not.
Methods: A Gaussian process regression soft sensor modeling method with real time integration learning is studied in depth by combining two local learning strategies, namely just-in-time learning (JITL) method and integrated learning method, and a multi-model weighted Gaussian process regression (MWGPR) soft sensor modeling method based on real-time integration learning is proposed in the paper. This soft sensing method was used to study the relationship between the viscosity of fermentation broth and the contamination in fermentation process. A soft-sensing model based on the viscosity of fermentation broth for predicting the signs of contamination is established.
Results: The validity of this method is verified by field data. The experimental results demonstrate that the soft sensing model proposed in this paper can effectively determine whether the fermentation broth is infected by hybrid bacteria.
Conclusions: The method proposed in this paper is innovative and practical so that field operators can issue early warning and take effective measures.
Keywords: Chlortetracycline fermentation; Gaussian process regression; contamination prediction; just-in-time learning; soft sensor modeling.