针对污水处理工艺中关于曝气系统中溶解氧对象非线性,实时误差大,普通PID控制器和神经网络控制器无法良好控制的问题,在分析和建立其模型的基础上设计了一种基于神经网络的溶解氧非线性PID智能控制器。上述控制器的比例,积分和微分参数采用关于误差的非线性函数的形式表示,并且在神经网络中训练比例,积分和微分神经元的权值,实现控制器参数的在线自适应调整来实现溶解氧非线性参数的控制。仿真结果表明,上述控制器有效的解决了传统PID对于非线性对象难以控制的问题,对外部干扰具有良好的抵抗性。能够满足污水处理过程中对于溶解氧的特殊控制要求。
A kind of nonlinear PID - like neuron network controller for dissolved oxygen in sewage treatment process is designed. The PID controller is used to adopt the nonlinear functions according to the error, and the weights of PID neurons are trained in the neural network to accomplish the online self - adaption of the controller and the non - linear parameter in dissolved oxygen as well is controlled. The result of simulation shows that the controller can effectively control the nonlinear objects ,which are hardly controlled by traditional PID controllers. And it has excellent performance to external disturbance. It can cater to the special requirements in the sewage treatment process.