基于BP神经网络的火电厂水质调节自适应控制的仿真研究

Simulation study of the adaptive control algorithm based on BP-NN for water quality adjusting in power plant

针对火电厂锅炉水质调节过程的大时滞时变特性,常规控制算法控制效果不好的问题,本文提出了基于BP神经网络的Smith-PID鲁棒自适应控制算法,利用BP神经网络的任意非线性表达能力和很强的自学习能力,在线自学习整定PID参数,被控对象不需要精确辩识,控制器参数跟踪被控对象自适应调整,克服了常规PID算法不适用于大时滞过程控制和常规Smith预估补偿控制对模型不确定性敏感的缺陷。MATLAB仿真表明,本文控制算法的静态特性、动态品质良好,鲁棒性强。

Aiming at the large time-delay and mutable characteristic of the adding-drug disposing process for the quality adjusting of boiler-water in power plant and the general control algorithm is of bad control effect, a robust adaptive Smith-PID control algorithm based on BP neural network is presented. In this algorithm, the application of the discretional nonlinear expression had powerful selfstudy capability of BP-NN to PID parameter tuning, it needn＇t to be identified the controlled-object accurately, and the controller parameter can change along with the change of controlled object characteristic. So it overcome the disfigurement of the general PID algorithm, which is misfit the control of the large time-delay and mutable process, and the general Smith-predictor control, which is impressionable to model error. MATLAB simulation showed that the controlling algorithm is of the better static characters, preferable dynamic quality and strong robustness.