基于改进混沌粒子群优化算法的热连轧厚度控制的研究

发布时间：2018-03-16 02:02

本文选题：厚度控制　切入点：混沌　出处：《华北理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：板带钢的质量水平高低在很大程度上取决于热连轧厚度控制系统,随着国内经济的发展,钢材市场中出现了前所未有的激烈竞争,所以更多的企业注重通过提高产品的质量来满足客户的需求。但是近年来我国板带钢产能出现过剩,再加上板带钢产品的厚度不达标等原因出现了很多质量问题,这些状况严重影响了企业的正常发展,所以追求更高的厚度精度就成为该领域的关注焦点。在科技迅速发展的今天,以传统控制理论为依据的带钢生产中,有很多困难无法处理,已不能满足如今高品质要求的市场和工业建筑对钢材的大量需求,新的控制理论和方法由此诞生,以弥补传统控制方法的不足,满足市场的需求。针对标准粒子群算法在寻优过程中收敛速度慢、易陷入局部最优的缺点,课题提出了一种依据种群早熟收敛程度和个体适应值来调整惯性权重的混沌粒子群优化算法,充分利用混沌优化与粒子群优化的搜索特征,训练神经网络的权值阈值,解决了BP神经网络收敛速度慢和易陷入局部极值的问题,仿真结果表明该算法不仅提高了误差精度,同时也加快了训练收敛的速度。随后将PID控制规律融进优化后的神经网络中,利用改进的混沌粒子群神经网络整定PID参数用于厚度控制系统,实现神经网络和PID控制规律的本质结合,从而构造出了一种智能PID控制器,并就系统快速性、抗干扰性和鲁棒性与传统PID控制器进行综合分析比较,仿真结果表明,前者的性能明显优于后者,改善了传统PID控制器在人为设定后就很难满足时变性要求的缺陷,该控制方法稳定可行。
[Abstract]:The quality level of strip steel depends to a great extent on the thickness control system of hot strip rolling. With the development of domestic economy, there has been unprecedented fierce competition in the steel market. Therefore, more enterprises pay attention to meeting the needs of customers by improving the quality of their products. However, in recent years, there has been a lot of quality problems caused by the overcapacity of steel plates and strips in China, and the fact that the thickness of steel products is not up to standard. These conditions have seriously affected the normal development of enterprises, so the pursuit of higher thickness accuracy has become the focus of attention in this field. With the rapid development of science and technology, in the production of strip steel based on traditional control theory, There were many difficulties that could not be dealt with, which could no longer meet the large demand for steel in the market and industrial buildings, which now required high quality, and new control theories and methods were born to make up for the shortcomings of the traditional control methods. To meet the needs of the market, the standard particle swarm optimization algorithm in the process of convergence is slow, easy to fall into local optimum, In this paper, a chaotic particle swarm optimization algorithm, which adjusts inertia weight according to the degree of premature convergence and individual fitness of population, is proposed to train the weight threshold of neural network by making full use of the search features of chaos optimization and particle swarm optimization. The problem of slow convergence speed and easy to fall into local extremum of BP neural network is solved. The simulation results show that the algorithm not only improves the accuracy of error, but also improves the accuracy of BP neural network. At the same time, the speed of training convergence is accelerated. Then the PID control law is incorporated into the optimized neural network, and the improved chaotic particle swarm optimization neural network is used to adjust the PID parameters for the thickness control system. The essential combination of neural network and PID control law is realized, and an intelligent PID controller is constructed. The system's rapidity, anti-jamming and robustness are analyzed and compared with the traditional PID controller. The simulation results show that, The performance of the former is obviously better than that of the latter, which improves the defect that the traditional PID controller is difficult to meet the time-varying requirement after artificial setting, and the control method is stable and feasible.
【学位授予单位】：华北理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG335.5;TP18

【参考文献】