【作者】 张丽；

【导师】 苏义鑫；

【作者基本信息】 武汉理工大学 ， 控制科学与工程， 2011， 硕士

【摘要】 主动磁力轴承作为一种优秀的机电综合体,它具有许多老式的接触式轴承所不具备的优点,比如没有摩擦,故没有磨损,无需在轴承转子和定子之间涂润滑剂,因此转子运动更快,使用寿命更长。正因为这些优点,主动磁力轴承受到工业领域比如轴承行业以及学术领域的广泛关注和热议。但由于磁力轴承本身固有的特性,如不稳定性、参数不确定性、模型存在非线性等。在过往的研究中发现,采用传统的PID控制器无法达到理想的控制要求。需将新的算法加入其中进行分析研究。本文主要针对磁力轴承中的单自由度磁力轴承进行讨论分析。在对磁力轴承电磁力和受力问题的分析后,对磁力轴承非线性特性进行建模。然后在传统PID闭环控制的基础上,加入BP神经网络算法和支持向量机算法对PID的控制参数进行调整,通过仿真实验对比两种算法的控制效果。文章介绍了神经网络的学习规则,利用神经网络的高度非线性映射能力,分析设计了BP神经网络PID控制器,仿真结果表明,BP整定PID控制可以有效地减小超调,增加转子的起浮位置。但因为神经网络存在局部极小、易出现过拟合等问题,随着隐含层节点数目的增加,控制性能反而变差。为避免神经网络的缺点,文章提出了基于支持向量机的磁力轴承PID控制。在对支持向量机的基本理论及其回归算法进行了详细介绍后,首先利用支持向量机能逼近任意非线性函数的特点,在传统PID闭环控制的前提下,对磁力轴承的非线性系统进行辨识。然后推导出基于支持向量机的PID控制器算法,结合辨识模型,利用Simulink中的M函数和SVM工具箱实现基于支持向量机PID控制的磁力轴承控制系统的仿真实验。将其与BP神经网络整定PID控制和传统PID控制相比较,仿真结果表明,基于支持向量机的自适应PID控制器的控制效果更好,不仅可以使磁力轴承在更宽范围内起浮,而且调节时间快。更多还原

【Abstract】 As the excellent electrical and mechanical complex, Active Magnetic Bearings have many advantages such as no friction, no frays and dispensing with lubricate compared with traditional bearing. Because of these advantages, active magnetic bearings have been attracting great attention.However, due to the inherent characteristics of magnetic bearings, such as instability, parameter uncertainty, the model is nonlinear and so on. In the previous study, the usual idea is to get the linearization model at the balance point of the magnetic bearing, then to design the traditional PID controller based on linear theory. Found that using the traditional PID controller can not achieve the desire of the control requirements. So we need to find new control algorithm.In this article, the reasonable mathematic model was built up based on the study of the single freedom magnetic bearings, and the model of the control system was given too. In the base of traditional PID control, the BP neural network algorithm and the support vector machine algorithm are added to adjust the parameters of the PID controller. Then compare the two methods by simulation.The article introduces the neural network learning rules, the height nonlinear mapping ability of the neural network, analysis and design the BP-PID controller. Simulation results show that, this kind of control strategy can reduce the overshoot, and make the rotor of the magnetic bearings stabilize in a large range. Because neural network exists local minimum, and is prone to over-fitting, with the increase in the number of hidden layer node, the control performance becomes worse.To avoid the shortcomings of the neural networks, the PID controller based on support vector machine is designed. After introducing the basic theory and the regression algorithm in detail, first, according to the characteristics of support vector machine can approach any nonlinear function, in the traditional PID closed-loop control, identify the magnetic bearing system. Then, combined with the identification model, using M function and toolbox of SVM, design the PID control algorithm based on support vector machine. Compared with BP-PID controller and traditional PID controller, the simulation results indicates that the control characteristics are much superior, it makes the rotor of the magnetic bearings stabilize in a large range and shortens the adjustment time.更多还原