基于RBF人工神经网络的下游常水位自适应渠道输水控制研究

针对传统渠道输水PID控制方法响应速度慢、超调量大、参数不能在线自调整的不足,根据RBF神经网络和渠道输水特点,提出了将传统渠道下游常水位输水PID控制和RBF人工神经网络结合的输水控制方法,使输水控制具有自学习、自适应、容错性和鲁棒性。推导了RBF网络整定PID输水控制调节器的算法。仿真结果表明,基于RBF网络的PID输水控制方法,能够通过不断学习自动调整控制参数,使输水控制过程超调量小、响应速度快,具有不需要特意选择或计算控制参数的特点。因此,基于RBF神经网络的参数非线性PID控制更适合进行渠道输水这样高度非线性系统的实时控制过程。

Research of self-adapting canal downstream constant level control based on RBF neural network

The classic PID canal control models have some disadvantages such as slow response speed, larger over-shot and non self-adjusting on-line. So, a method combining normal PID and RBF neural net-work method is presented according to the characteristic of canal transmitting water. It has merits of self- studying,self-adapting,fault tolerance and robustness. The algorithms of RBF Neural Network PID Control of canal transmitting water are deduced. Through hydraulic simulating, the results show that in the process of RBF PID control,the method can adjust the parameters to optimal state, the water level overshoot is small and the response is quick. It can be required to choose computer parameters complexly,and can be used for the real-time control of nonlinear system such as canal transmitting water. So the RBF neu- ral network PID is a more suitable model for real-time nonlinear canal control than classic one.