Study On The Superheated Steam Temperature Control System Based On Fuzzy-Neural Network

Currently,the superheated steam temperature control system in most power plants still adopts the conventional PID cascade control scheme. The larger capacity and the higher parameters the thermal power Turbo-generator unit has, the worse control performance the conventional PID control scheme gets under different running conditions, because the superheated steam temperature is a complex controlled object with large delay, large inertia and time-varying as well as nonlinear property.A new superheated steam temperature control system design scheme is proposed in this paper, which tries to combine the advantages of both intelligent control and conventional control. This design not only keeps the conventional cascade control's basic structure with its anti-jamming function, but also adds the good dynamic performance to the system by means of main controller design based on Fuzzy-Neural Network control. The main controller is composed of a Fuzzy-Neural Network controller and a subsystem of intelligent integral. Moreover, the fuzzy controller's suitable fuzzy rules are auto-generated after the neural network learns from the training data; the subsystem of intelligent integral is able to work at the just time when it is needed.At last, large numbers of simulations are taken using the MATLAB toolbox, and compared with the conventional PID control scheme. The results show that: the system this paper designed obviously has a better control quality than PID does, so it has a promising future in the thermal process control.