摘要
为提升掘进机摆动截割破岩时BP神经网络PID控制的自适应性,提出了改进粒子群优化BP神经网络PID控制方法.分析了悬臂式掘进机水平摆动截割系统的工作原理,构建其系统的传递函数,设计了BP神经网络PID控制器,确定了BP神经网络的结构和参数,推导出了BP神经网络权值和输出层阈值梯度,引入Tent混沌映射来改进粒子群算法,优化了BP神经网络的权值,以阶跃函数作为输入信号,仿真研究该算法的性能.结果表明,与未优化的BP神经网络PID控制器相比,优化后的BP神经网络PID控制器,能够快速准确地跟踪阶跃响应.
Abstract
This paper aims to improve the adaptability of BP neural network PID control of the road-header,while swinging and cutting the rocks,and proposes an improved particle swarm optimization of BP neural network PID control method.The study operates by analyzing the working principle of the hori-zontal swing cutter system of the cantilever roadheader;constructing the transfer function of the system;designing the BP neural network PID controller;determining the structure and parameters of the BP neu-ral network;deriving the BP neural network weights and the threshold gradient of the output layer;intro-ducing Tent chaotic mapping to improve the particle swarm algorithm;optimizing the weights of the BP neural network;and simulating and studying the performance of this algorithm by using step function as the input signal.The results show that compared with the unoptimized BP neural network PID controller,the optimized BP neural network PID controller can track the step response quickly and accurately.
维普
万方数据