渠道系统神经网络控制运行研究

渠系运行中存在非线性、大滞后及时变性的特点。利用神经网络处理复杂非行线性、不确知系统的能力,设计基于PID神经网络的水位反馈控制器,它既具有常规PID控制器结构简单、可靠性高的优点,又具有神经网络自学习、自适应的能力,实现了控制器参数整定不依赖于渠道系统数学模型,且能依据渠系动态信息适时调整。结合渠道水力学特性建立渠道自动化运行数学模型,并进行了计算机数值仿真。结果表明,采用PID神经网络控制器的渠道运行系统,实现了控制参数自适应调整,且动、静态性能较常规PID控制有明显改善。

Study on Canal System Operation Controlled by PID Neural Network

Canal system has the characteristics of non-linearity,long time lag of dynamic response and time variation.Neural network is just good at dealing with nonlinear and uncertain systems,so a water feedback controller based on PID neural network was developed.It has both the merits of PID rule and neural network such as simpleness and high reliability,self-learning and selfadaptive ability.Furthermore,the control parameters can be integrated not depending on the canal mathematical model and can be adjusted according to the real-time information of the canal system.Based on canal hydraulics characteristics,a mathematical model of canal system was constituted and the computer simulation was carried out.The simulation results prove that the performance of canal system operation controlled by PID neural network control is better than that controlled by conventional PID control.