摘要
磨煤机制粉系统具有非线性、时滞性大等特点,而且其输入量和输出量之间高度耦合,采用常规的控制方法难以实现良好的效果。为此提出模糊控制与神经网络结合的结构,采用误差反传优化网络权值,用于磨煤机制粉系统的控制。首先采用模糊算法对神经网络的输入值进行预处理,然后采用误差反传算法调整网络的权值,实现了PID控制器参数的自适应调整。仿真实验表明,该方法能解决耦合性、时滞性问题,超调量小,跟踪效果好,具有良好的鲁棒性和适应性。
The pulverizing system of coal mill has the characteristics of nonlinearity and large time delay, and its input and output are highly coupled. It is difficult to achieve good results by using conventional control methods. Therefore, this paper proposes a structure combining fuzzy control and neural network, the weights of network was optimized by using the error back propagation, which is used for the control of pulverizing system of coal mill. Firstly, fuzzy algorithm is used to preprocess the input value of neural network, then the back propagation algorithm is used to adjust the weight of the network, adaptive adjustment of PID controller parameters is realized. Simulation results show that this method can solve coupling and time delay problems, which has small overshoot, good tracking effect and good robustness and adaptability.
引文
[1]蔡宇宁,杜平.PID智能控制在磨粉系统中的实际应用[J].有色矿冶,2011,27(3):84-85,97.
[2]蔡改贫,许琴,曾艳祥,等.预磨机磨矿系统的IMC-PID串联解耦控制[J].北京工业大学学报,2016,42(1):35-41.
[3]刘振东,王建民,杨刚.基于改进BP-PID控制的磨矿回路控制系统的研究[J].矿业研究与开发,2018,38(7):99-103.
[4]孙杰,韩艳,段勇,等.基于改进的PSO算法的球磨机PID神经网络控制系统[J].工矿自动化,2011,37(5):59-62.
[5]韩良云,陆金桂.神经网络球磨机出力软测量模型的建立[J].动力工程学报,2015,35(11):901-905,933.
[6]程启明,程尹曼,汪明媚,等.球磨机混合优化前向神经网络PID解耦控制系统[J].电力系统及其自动化学报,2010,22(2):54-59.
[7]张会娟,刘杰,陈红梅,等.基于模糊神经网络PID的小麦制粉能耗控制[J].粮食与油脂,2019,32(3):27-30.
[8]朱丽娟.球磨机对象控制中模糊径向基函数神经网络的PID控制分析[J].现代电子技术,2015,38(24):56-58,61.
[9]孙林,杨琳霞,吴耀华.火电厂钢球磨煤机专家模糊PID解耦控制方法研究[J].热力发电,2011,40(7):54-57.
[10]李琳娜,蒋彬.基于模糊控制的智能变频空调控制系统研究[J].廊坊师范学院学报(自然科学版),2011,11(6):49-52.