基于小波神经网络模型的预测控制二容水箱

With coming of the era of economy, improvement of production technology is becoming more and more urgent, however, the scene of the industrial complex and changeable makes experiment for any small mistakes, not only can cause damage to equipment, more can cause waste in time, in the face of these problems, can simulate the most real industrial object is key to the emergence of four water level control system solved the problem to a great extent, especially the nonlinear and coupling characteristics, makes the nonlinear predictive control in the control system also has the object to study, the stand or fall of nonlinear approximation ability of wavelet neural network, is often critical link in model accurately or not, so in this paper, the wavelet neural network to optimize the parameters of the early and in the process of optimization, and put forward the strategy, a linear gradient factor can make the variance decreases trends, balancing algorithm's role, then to establish exact model combined with dynamic matrix control strategy, the more water to study the nonlinear and control. Face more water coupling characteristics, the use of neural network decoupling technology to deal with this problem, the main principle of neural network identification is combined with feedforward decoupling theory, the decoupling channel, neural network structure and parameters for the relevant design, gives the specific strategy of neural network decoupling, through the simulation results show that the control strategy in this paper, design of the control effect is relatively good, and the detected after joining fixed interference sources of stability and robustness is also good, in order to solve the problems of the nonlinear and coupling provided a feasibility.