摘要
由于废水成分和流量波动大,导致压力频繁变化,同时多泵并联或串联运行时压力调节相互影响,导致废水处理中给水泵出口压力控制难度非常大。传统给水泵出口压力控制方法主要依赖于比例-积分-微分(Proportional-Integral-Derivative control,PID)控制器对出口压力进行调节,忽略了出口压力数据的非线性、时变特点,导致控制过程超调量较高,稳定性不够理想。对此,提出基于改进遗传算法的废水处理中给水泵出口压力模糊控制方法。结合梯度提升树(Gradient Boosting Decision Tree,GBDT),通过迭代构建多个决策树,对出口压力数据的非线性、时变特征进行挖掘,对废水处理中给水泵出口压力进行预测。将模糊控制法与PID算法进行结合,以压力误差作为输入变量,对模糊控制器进行设计。将模糊控制器的关键参数进行编码,并将参数编码结果作为遗传算法的个体,采用改进后的算法对个体进行寻优操作,从而实现控制器参数优化。在实验中,对提出的方法进行了控制稳定性的检验。结果表明,采用提出的方法对给水泵出口压力进行控制时,算法的超调量均值为2.8%,具备较为理想的控制稳定性。
Due to the large fluctuations in wastewater composition and flow rate,pressure changes frequently,and the mutual influence of pressure regulation when multiple pumps are operated in parallel or series,making it very difficult to control the outlet pressure of the feedwater pump in wastewater treatment.The traditional control method for outlet pressure of feedwater pumps mainly relies on proportional integral derivative(PID)controllers to regulate the outlet pressure,ignoring the nonlinear and time-varying characteristics of outlet pressure data,resulting in high overshoot and unsatisfactory stability in the control process.A fuzzy control method for the outlet pressure of the feedwater pump in wastewater treatment based on improved genetic algorithm is proposed.Combining gradient boosting decision tree(GBDT),multiple decision trees are iteratively constructed to mine the nonlinear and time-varying characteristics of outlet pressure data and predict the outlet pressure of water pumps in wastewater treatment.Combining fuzzy control method with PID algorithm,using pressure error as input variable,design a fuzzy controller.Encode the key parameters of the fuzzy controller and use the encoded results as individuals for the genetic algorithm.Use the improved algorithm to perform optimization operations on the individuals,thereby achieving controller parameter optimization.In the experiment,the stability of the proposed method was tested for control.The final test results show that when using the proposed method to control the outlet pressure of the feedwater pump,the average overshoot of the algorithm is 2.8%,which has a relatively ideal control stability.
作者
范巍
FAN Wei(SUEZ Environment technology(Beijing)Co,.Ltd.,Beijing 100026,China)
出处
《国外电子测量技术》
2025年第5期193-199,共7页
Foreign Electronic Measurement Technology
关键词
改进遗传算法
废水处理
给水泵
出口压力
模糊控制
improved genetic algorithm
wastewater treatment
feed pump
outlet pressure
fuzzy control
