An Application Research Of Fuzzy Neural Network Used In Superheated Temperature Control Of Boiler

The superheated steam temperature in most thermal power boiler is a charged part which has large time delay, large inertia, and it also has strong parameter variability and nonlinearity characteristics. Therefore, it is difficult to control the steam temperature effectively. When disturbed variable is happened, whether adjust the control system and keep the steam temperature error in range or not is important to the control system. It affects the safety and economy of thermal power operation.This paper presents an adaptive fuzzy neural network superheated temperature control scheme based on studying adaptive neural fuzzy inference system (ANFIS) and fuzzy control deeply, and discusses the response characteristics of superheated steam temperature using Matlab/Simulink.The content of this researched thesis mainly included the following aspects:(1)This paper described the research significance of the steam temperature, introduced the research present situation and developing trend. We established the paper's starting point based on some problems existed in the control strategy nowadays.(2)Introduced static and dynamic characteristics of superheated steam temperature briefly, described methods of establishing steam temperature mathematical model, and highlighted steam temperature cascade control system.(3)Introduced the theories of fuzzy control and neural network control, and analyzed the design steps of fuzzy controller and learning algorithm of neural network emphatically.(4)Described fuzzy neural network control strategy briefly, and detailed learning algorithm of fuzzy neural network, designed steam temperature cascade control system whose main controller is adopted a fuzzy neural network controller (FNNC), the simulation shows that this control method is more effective than traditional PID control.(5)Elaborated ANFIS neural network, introduced the structure and learning algorithm of ANFIS, designed steam temperature cascade control system. The main controller is constituted of a fuzzy controller based on ANFIS neural network and an intelligent proportional integral one by parallel. Finally, it is applied in 600MW boiler and is simulated, the simulation result shows that this system has good control effect and certain engineering applications.