基于广义预测控制算法的网络控制系统时延补偿研究

发布时间：2018-06-26 01:33

本文选题：网络控制系统 + PSO-BP　；参考：《沈阳工业大学》2017年硕士论文

【摘要】：随着控制理论、通信技术、计算机的发展,网络控制系统在工业控制领域得到越来越广泛的应用。网络控制系统各个节点的数据传输和交换是通过通信网络进行的,必然会产生随机网络时延的问题,影响系统的控制性能。传统控制的研究方法已经不能满足当前网络控制系统研究的需要。为了应对网络控制系统带来的挑战,设计出适用于实际控制系统的控制策略提高系统的控制性能是亟待解决的事情。研究内容包括时延预测和控制器设计两部分。首先,介绍网络控制系统的背景、研究现状。总结国内外学者的研究成果。其次,介绍网络控制系统的基础知识。分析了网络控制系统的基本结构和网络时延的组成,说明时延研究的必要性。利用TrueTime搭建仿真平台。推导了时延和采样周期二者之间对系统稳定性的影响。然后,对粒子群优化算法和BP神经网络算法的基本原理进行介绍,提出改进PSO-BP复合算法建立预测模型,为时延的预测打下基础。从实验仿真平台中获取到历史时延数据,进行数据处理并进行仿真验证,获得很好地预测结果。最后,对预测控制算法进行了分析。根据工业生产过程遇到的问题,对算法进行一步一步的改进,提出效果更好的PI型隐式GPC控制算法。对PI型隐式GPC的稳定性进行了分析推理。提出了基于改进PSO-BP的时延预测方案和PI型隐式GPC控制算法的控制器设计方案。仿真结果表明本文提出的设计方案具有良好的补偿效果。
[Abstract]:With the development of control theory, communication technology and computer, networked control system has been more and more widely used in industrial control field. The data transmission and exchange of each node in the network control system is carried out through the communication network, which will inevitably cause the problem of random network delay and affect the control performance of the system. The traditional control research method can not meet the needs of the current networked control system research. In order to meet the challenge of networked control system, it is urgent to design a control strategy suitable for practical control system to improve the control performance of the system. The research includes two parts: time delay prediction and controller design. Firstly, the background and research status of networked control system are introduced. Summarize the research results of domestic and foreign scholars. Secondly, the basic knowledge of network control system is introduced. The basic structure of networked control system and the composition of network delay are analyzed, and the necessity of studying delay is explained. Using TrueTime to build simulation platform. The influence of time delay and sampling period on the stability of the system is deduced. Then, the basic principles of particle swarm optimization and BP neural network algorithm are introduced, and the improved PSO-BP compound algorithm is proposed to build a prediction model, which lays a foundation for the prediction of time delay. The historical delay data are obtained from the experimental simulation platform, and the data are processed and verified by simulation, and good prediction results are obtained. Finally, the predictive control algorithm is analyzed. According to the problems encountered in the industrial production process, the algorithm is improved step by step, and a better Pi implicit GPC control algorithm is proposed. The stability of Pi implicit GPC is analyzed and inferred. A time delay prediction scheme based on improved PSO-BP and a controller design scheme based on Pi implicit GPC control algorithm are proposed. Simulation results show that the proposed design has a good compensation effect.
【学位授予单位】：沈阳工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP273

【参考文献】