Fuzzy Neural Network Controller Based on Improved Genetic Algorithm for Temperature of Glass Tempering and Annealing Process

The temperature control of the glass tempering and annealing process has characteristics of time-varying parameters, nonlinear and big lag. It is difficult to meet the expected control effect with the common control method. To solve this problem, this paper puts forward a kind of fuzzy neural network controller optimized by genetic algorithm. First, it uses neural network to construct fuzzy logic system according to the structure equivalence rule, thus the optimization of fuzzy control rules and membership functions can be realized by finding the weight value of the neural network. Then, it uses the improved genetic algorithm to find the global optimum weighted factors with a high speed so to improve the performance of the controller. The simulation results show that the optimized fuzzy neural network controller can obtain an excellent control performance for the nonlinearity system with time- varying parameters and lag.