The Real-time Simulation And Research On Neural Pid Temperature Control System

The PID control method is one.of the traditional control methods and gets good effects under many application situations. But with the increase in complexity of manufacture technics and demands of industrial process performance, the conventional PID control can not meet the requirement of closed loop optimized control. Neural network, as one of modern information process technologies, has some advantages in many applications. Neural Network PID control technology in which plays a very important role, and will become one of the focus of the research and application of technology.In this paper, the current study, more mature neural network PID control are summarized. The basic principle of neural network PID control is expounded. Some of the typical neural network PID controllers (including the Single Neuronal PID Controller, the Neural Network PID Controller and the liked PID-NN Controller) are studied by MATLAB/Simulation tools in order to show their characteristic and applicability.Temperature control system is the large lagging, strong coupling, slow time-varying and nonlinear complex systems. In the temperature control system, controlled object (temperature control box) has parameter uncertainty and time delay characteristics. Precise mathematical model is difficult to be established. In this paper, through temperature control system about the controlled object digital simulation and real-time simulation study, the characteristic of the neural PID controllers is compared with the conventional PID control. By Analysis of the traditional PID controller and the neural network PID controller advantages and shortcoming, corresponding improvement programmes were proposed.