The Research And Application Of Wavelet Neural Network In Multi Water Tank

In the process of industrial production,it often has the characteristics of non-linear,time-varying,strong coupling,in process control,PID alone is often difficult to achieve stable effect at the same time,the sensor of industrial production can not guarantee the accuracy of data,it will promote the application of intelligent algorithm in PID control.Wavelet analysis has the strong approximation ability,and the neural networks has a strong self-learning ability,combining the two in tuning of PID parameters using wavelet neural network.By means of an experimental platform similar to an industrial device,it verifies the ability of wavelet neural network in the process of industrial production.In this paper,based on the typical four tank,the wavelet neural network algorithm based on PID control is studied.The main work is focused on the adjustment of PID parameters by wavelet neural network.In this paper,a PID controller and a wavelet neural network PID controller are designed for the mathematical model of a single tank and a double water tank.The simulation results show that the Matlab neural network can adjust the PID parameters better.In this paper,a fuzzy decoupling controller is designed for four tank with strong coupling.The conclusion is reached that the water level is stable after decoupling.In the CS4000 experiment device,the operation process by using configuration software to monitor the whole water tank,the PC real-time data validation in the setting effect of double tank and four tank wavelet neural network PID controller to adjust the parameters of PID controller is better than that of PID itself,the parameters of wavelet neural network tuning after the water level reached stable faster,and small oscillation,small overshoot.Better control effect of PID controller based on wavelet neural network.The wavelet neural network proposed in this paper provides a solution for the complex control of four tank water level system.