Research on the monitoring system method of GA-BP network soft-sensing

Combination neural network and genetic algorithms consisting of GA-BP network, which builds soft-sensor model. GA-BP network for modeling the core soft-sensing method can avoid the use of BP algorithm alone into the problem of local minimum, also speed up the network convergence rate. The realization of monitoring system is firstly the trained GA-BP soft-sensing model used in computer control system by program. For fermentation process of germ concentration as a example, collecting the process data of industrial field as the input of soft-sensing, and germ concentration as output, so that it can be achieved germ concentration on-line estimate, thereby enhancing the production of glutamic acid, reducing production cost increasing economic efficiency.