具有神经网络前馈控制器的污水处理模糊控制系统设计

摘要

A novel method that combined fuzzy controller and the feed forward controller for a sewage treatment plant is proposed, a nonlinear, time-delay, highly interfered system. Except this, how to design a feed forward controller for the sewage treatment plant is also researched in this paper. A new feed forward controller based on back propagation （bp） neural network is designed by the method which didn＇t need mathematical models of control and interference channel. The control system is simulated and simulation results shows that the method is more effective than single fuzzy control method with higher performance in elimination of interference and reducing stable error. Simulation of the feed forward controller shows that the feed forward controller based on bp networks could compensate interference in the fuzzy control system, and it＇s valuable especially when mathematical models of control and interfere channels is unknown, though it＇s performance is poorer than normal feed forward controller.%为解决污水处理过程中非线性、大时滞及干扰严重的问题，提出一种将模糊控制器和前馈控制器相接合的控制方法。其中在建立前馈控制器时，提出了使用神经网络建立前馈控制器的新方法。使用这种方法建立前馈控制器无需建立控制系统控制通道和干扰通道的数学模型。也不要求干扰是可以测量的。在MATLAB环境下对控制系统进行仿真，结果表明，将模糊控制和前馈控制结合可以有效克服干扰，提高系统的稳态精度；基于神经网络的前馈控制器可以有效补偿污水处理过程中的干扰，具有实际意义。