Research Of The Nonlinear Identification Algorithm For The Thermal Process System

In the process of thermal system identification,prediction and analysis,identifying the model of system is usually necessary.Accurately identifying the system will help to control the system more effectively.However,due to the models of the thermal system are nonlinear,it is difficult to establish an accurate mathematical model to describe the system.With the development of intelligent algorithms and statistical learning theory,using intelligent algorithms to fit the model of the system has gradually become a popular way for system identification.In this paper,after studying the characteristics of the thermal process in power plant with system engineering 6?implementation methods,we proposed new identification algorithm based on the neural network and support vector regression to identify the nonlinear system and simulation.For the system identification algorithm based on neural network,the papers presents a neural network based on fireworks algorithm optimization,and simulate this algorithm on steam pressure and stem temperature in thermal process control model.The results show that our algorithm can identify the system and anti-noise better.For the system identification algorithm based on support vector regression,the papers proposed a fuzzy support vector regression based on the distance membership function and simulation.The results show that compared with the traditional support vector regression,the fuzzy support vector regression can be fitting the system better and anti-noise at the same time.The papers developed a software for system identification based on the above Neural Network and SVM.which can automatically generate simulated input and output data for common thermal process model and implement system identification employing the real input and output data.