基于RBF网络的水厂投药预测控制

基于神经网络的智能预测控制具有很强的自学习和自适应能力,对于大滞后、非线性的复杂系统有较好的控制效果。针对水处理中混凝投药过程的特点并结合国内目前投药控制的现状,提出了一种带有前馈补偿的RBF神经网络预测控制新方法。该方法利用神经网络建立投药量的预测模型,然后用出水浊度与设定值间的预测偏差构成闭环控制。通过实时的在线滚动优化,实现了投药量的最优投加。仿真试验表明,出水浊度保持稳定,所需矾耗减少,控制效果明显。

Predictive Control of Water Treatment Plant Dosage Based on RBF Network

Intelligent predictive control based on neural network has good ability of self-learning and self-adapting,therefore,it has better control effect for nonlinear large time-delay complex system.By analyzing the characters of coagulant dose process and the status of domestic dosage control,a new method of predictive control based on RBF network with forward compensation is proposed.A predicting model is established by neural network,and then a closed-loop control system is constructed by the predictive deviation between turbidity of outlet water and the set-value.Through real-time on-line rolling optimization,optimal chemical dose rate is realized.The simulation of coagulant dose process shows that steady turbidity of outlet water is kept,and coagulant consumption is reduced,the control effect is perfect.