基于改进粒子群优化的双闭环自愈调控算法研究 (2)

硕 士 学 位 论 文 中文题目： 基于改进粒子群优化的双闭环自愈调控算法研究 The Research of Double Close Loop Self-healing System 英文题目： Based on The Improved Particle Swarm Optimizer 专 业： 计算机应用技术 摘要 摘要 近年来，工业生产自动化不断提高，作为机电设备中核心的动力设备，电机在工业生产的各个领域应用也越加广泛。在工业生产过程，电机发生故障不仅会对生产进度和质量造成严重的影响，而且对工人的人身财产安全造成巨大的危害与威胁，因此确保电机正常工作就显得极其重要。在工作过程中，因为外界或者自身的因素变化，电机运行状态会产生变化，例如工作产热引起的工作温度变化等。温度升高，电流升高，若保持负载能力不变，长期运转后，电机会因温度过高而烧毁。我们称这种“带病”状态为“亚健康”状态。国内外对此方面的研究一直非常重视，如何做到预防故障发生及使电机从“带病”工作状态恢复到正常状态一直都是非常重要的课题。 在阅读分析大量关于故障检测及控制的方法之后，对传统的故障控制算法及故障状态评价进行改进，本文提出一种针对电机运行时“亚健康”状态的双闭环自愈调控算法。本文算法在传统控制闭环基础上添加自愈闭环，形成包含控制环和自愈环的双闭环结构。控制环采用PID控制，对其参数采用粒子群算法进行优化整定，并针对粒子群算法容易陷入局部最优的特点，对其惯性参数和学习因子进行改进，得到改进的粒子群算法。自愈环对无刷直流电机的电流采用均方根转化的方法，对三相电流进行转换处理，结合“亚健康”因素与电机运行状态的关系设计划分函数。引入健康度的概念，对直流电机的健康等级进行划分，对直流电机的运行做出干预决策。算法是针对电机工作时，因外界或自身条件变化导致的电机运行状态变化所产生的潜在故障做出自愈调控，使之自行判断健康等级，并作出降载的操作，降低出现潜在故障的可能性。 在MATLAB平台中的SIMULINK下进行仿真实验，采用无刷直流电机为实验模型，对无刷直流电机的工作状态进行模拟，验证算法有效性。实验结果证明，在电机工作状态发生变化时，电机能在算法的控制指导下，自行完成降载，脱离“亚健康”状态。 关键词：双闭环自愈调控；亚健康；改进的粒子群算法；健康度； I Abstract Abstract In recent years, the automation of industrial production improve continually. The motor, as a core dynamic facility of machinery and electronic plant system, has been much more widely used in the industrial production in various fields. In the whole process of industrial production, the failure of the motor will not only cause serious impact to production schedule and quality, but also does great harm to personal safety property of workers. Thus, to ensure the normal working condition of the motor is extremely important. During the working process,  the running state of a motor will change because of some transformation of external factors or their own changes, such as the change in working temperature caused by heat production, etc. When temperature rising, the resistance of the load will reduce, which push the current up. If that kind of situation continues for a long period of time, the motor will burn down because of the high working temperature. We call this kind of working state as sub-health state. A large amount of research has been carried out at home and abroad. How to do prevent the fault and make the motor from sub-health state back to the normal state has been a very significance topic. After reading a lot about the fault detection and control methods, the paper puts forward a kind of double-close self-healing control algorithm to deal with the sub-health state of a running motor, which has improved a lot compared with the traditional control and evaluation algorithm. A self-healing ring is added into the traditional control ring, which constitutes a double closed structure. The PID controller is used in the control ring. The particle swarm optimization algorithm is adopted to adjust the parameter of PID controller. Aiming at the disadvantage of easily falling into local optimum, both the inertia weight and the learning factor are improved to raise the effectiveness of the algorithm to get the improved particle swarm optimization. The self-healing ring uses the RMS to transform the three phase current signals and combines the “sub-health” factor with the running state to design the partition function. The concept of heath degree is adopted to divide the health level and makes a decision for the running state of the DC motor. This algorithm is focus on that the motor can do some self-healing control to deal with the incipient fault caused by some transformation of external factors or their own during the running state. It can make the motor judge the health level itself and reduce the load, which can reduce the probability of the incipient fault. The SIMULINK simulation experiment is done in the platform of MATLAB. The brush-less DC motor is used to verify the effectiveness of the algorithm. The experimental result shows that when the motor running condition changes, the motor load can be reduced by the control algorithm itself and escape from the sub-health state. Key words: Double Close Loop Self-healing; Improved Particle Swarm Optimization; Improved SVM; Sub-health; Health Degree III 目录 目录 第1章 绪论 1 1.1 研究背景和研究意义 1 1.2 国内外研究现状 3 1.2.1 自愈控制算法的研究现状 3 1.2.2 “亚健康”理论的研究现状及发展 5 1.3 论文的主要研究工作 6 1.4 论文的结构安排 7 第2章 PSO算法、PID控制器及闭环控制的基础理论 9 2.1 群智能算法 9 2.1.1 蚁群优化算法 9 2.1.2 人工鱼群算法 9 2.1.3 群智能算法的优点 10 2.2 基础和标准粒子群优化算法 10 2.2.1 粒子群优化算法的建立 10 2.2.2 基本粒子群优化算法 11 2.2.3 标准粒子群优化算法 13 2.3 PID控制器简介 14 2.3.1 PID调节器模型 14 2.3.2 PID参数对控制系统的影响 14 2.4 闭环控制 19 2.4.1 反馈调节 19 2.4.2 闭环控制的优缺点 20 2.5 本章小结 20 第3章 基于改进粒子群算法对PID参数整定的研究 21 3.1 粒子群优化算法对PID参数整定的原理 21 3.1.1 控制系统性能指标 21 3.1.2 几类典型被控对象的数学模型 24 3.2 粒子群优化算法的改进 25 3.2.1 对惯性权重参数的改进 25 3.2.2 学习因子的改进 28 3.3 基于改进粒子群算法对PID参数整定的仿真研究 31 3.3.1 仿真的步骤与流程 31 3.3.2 仿真及对比试验 32 3.4 本章小结 34 第4章 双闭环自愈调控算法的研究 35 4.1 设备的“亚健康”状态 35 4.1.1 机械健康度（MHD）的定义 35 4.1.2 运行状态的机械健康度划分 36 4.2 双闭环控制算法 37 4.2.1 控制环 39 4.2.2 自愈环 39 4.3 划分函数的构建 40 4.