Fuzzy neural network inversion soft measurement system and method of key variables in lysine fermentation process

Abstract

The invention discloses a fuzzy neural network inversion soft measurement system and method of key variables in a lysine fermentation process. Online directly measurable quantity and indirectly measurable quantity needing off testing in a lysine fermentation process are determined. By use of the fuzzy neural network inversion soft measurement method, according to the model of the lysine fermentation process, the model of a soft sensor is established, thereby, establishing an inversion model of the soft sensor according to a method of solving a reverse function, and determining free parametersand constructing fuzzy neural network inversion by use of static state fuzzy neural network and a differentiator by training the dynamic fuzzy neural network. Thus, software sensor inversion is achieved. Finally, by serially connecting the fuzzy neural network to the lysine fermentation process, achieving online real-time soft measurement of cell concentration x1, substrate concentration x2 and product concentration x3. In this way, an online estimation problem that it is difficult to carry out real quantity measurement of key variables by use of a physical sensor in the lysine fermentation process can be effectively solved.

Description

technical field [0001] The invention belongs to the field of advanced control of microbial fermentation process, and in particular relates to a fuzzy neural network inverse measurement method of key variables in lysine fermentation process. Background technique [0002] In many industrial control occasions, there is a large class of such variables that are closely related to product quality and need to be strictly controlled. However, due to technical or economic reasons, it is currently difficult or impossible to detect directly through physical sensors. In order to solve the measurement of such variables Problems, soft sensor technology came into being. The so-called soft measurement is to select a group of directly measurable variables (i.e. auxiliary variables) that are closely related to the estimated variable (i.e. the measured or leading variable) and easy to measure according to a certain criterion. By constructing a certain functional relationship, use Computer sof...

AI Technical Summary

Problems solved by technology

These methods are complicated to operate, have long lag time, large measurement errors, and are easy to introduce artificial pollution, and the measurement accuracy is affected by factors such as cell death and measurement errors, which cannot reflect the current state of the fermentation process in time, and are difficult to meet the requirements of real-time dynamic regulation of lysine fermentation.

Images