The Application of Wavelet Neural Network in Adaptive Inverse Control of Hydro-turbine Governing System

Considering of the nonlinear, time-variable and non-minimum phase character and the easy variance of hydro-turbine governing system´s structure and parameters, a new adaptive inverse control method of hydro-turbine governing system based on the learning characteristic of neural network and the function approximation ability of the wavelet analysis is presented. It approximates the model and its inversion of plant by wavelet neural networks, and then through constructing an aim function of broad sense, a wavelet neural networks adaptive inverse law is put forward which is effective to the nonlinear non-minimum phase system. Theory and simulation to hydro-turbine governing system demonstrate that the control strategy can more effective improve the dynamic and stationary performance than those based on neural networks. It is showed the scheme is valid.