多算法优化融合控制球磨机制粉系统的研究  

Research on Multi-method Optimization and Fusion Control of Ball Mill Pulverizing System

作　　者：穆海芳[1,2] 韩君[1,2] 李明[1] MU Haifang;HAN Jun;LI Ming(School of Mechanical and Electronic Engineering,Suzhou University,Suzhou,Anhui Province 234000;Coal Mine Electrical Engineering Eechnology Research Center,Suzhou University,Suzhou,Anhui Province 234000)

机构地区：[1]宿州学院机械与电子工程学院,安徽宿州234000 [2]宿州学院煤矿电子工程技术中心,安徽宿州234000

出　　处：《楚雄师范学院学报》2020年第3期20-25,共6页Journal of Chuxiong Normal University

基　　金：安徽省科技重大专项(NO:18030901023)。

摘　　要：球磨机制粉系统具有的非线性、强耦合、时变性特点,使得实现其优良控制成为一个复杂的问题。在基于现有的控制方法基础上,提出多种算法混合优化的思想用于其生产过程的控制。首先建立控制器的结构,其次优化神经网络结构、优化反向传播算法、优化粒子群算法,完成PID算法的参数自适应调整,最后再对系统进行控制。通过对某球磨机系统的仿真实验表明,该方法较一般的控制方法具有超调小、跟踪快、鲁棒性强等特点,具有更好的控制品质。The pulverizing system of ball mill is nonlinear,strong coupling and time-varying,making it a complex problem to realize its good control.Based on the existing control methods,the idea of multi-method hybrid optimization is put forward for its production process control.Firstly,the structure of the controller is established;secondly,the structure of the neural network,the back-propagation algorithm and the particle swarm optimization is optimized to complete the PID algorithm parameters adaptive adjustment;and finally,the system is controlled.The simulation results of a ball mill system show that,compared with general control methods,this method has better control quality because it has the characteristics of small overshoot,fast tracking and strong robustness.

关 键 词：球磨机 模糊 神经网络 粒子群 

分 类 号：TG74[金属学及工艺—刀具与模具]