机电系统的故障诊断方法及应用研究

发布时间：2018-02-16 22:15

本文关键词： 故障诊断线性矩阵不等式滑模控制机电系统　出处：《湖南工业大学》2013年硕士论文　论文类型：学位论文

【摘要】：机电系统作为实现工业过程自动化的基本单元，是应用最为广泛的子系统。随着科学技术的进步以及现代化水平的提高，在工业现场使用的大量机电系统结构越来越复杂，自动化程度也越来越高，由于其使用的环境及工况等因素，使之发生故障的可能性越来越大，从而使对机电系统的故障诊断研究变得具有实际意义。本文以机电系统为工程背景，以风力发电系统和机器人等典型机电系统为具体对象，从提高对建模误差或干扰的鲁棒性，提高故障的诊断敏感性等几个方面着手，对机电系统的非线性故障诊断问题开展深入研究。其主要研究工作与结论如下：针对双馈风力发电系统空载并网过程中功率因数可能不为1的情况，建立了双馈风力发电系统空载并网的数学模型，并设计了积分型滑模变结构控制器，最后通过PSIM仿真与实验台实时实验证明了控制器的有效性。针对双馈风力发电机中的传感器故障，在考虑转速扰动的情况下，设计了一种基于滑模观测器与Luenberger观测器相结合的故障诊断方法，最后通过仿真实验证明其方法的有效性。针对一类机电系统的缓变故障，通过坐标变换将此机电系统模型变换为两个子系统，再利用滑模观测器和Luenberger观测器的互补性，分别对两个子系统设计了观测器，，实现了对微弱缓变故障的鲁棒诊断。在设计中，通过充分利用LMI算法，以达到放宽设计中关键参数的选择条件，从而对参数的求取更容易，使设计策略更具有实际可操作性。最后，将此故障诊断设计方法应用到单关节机械手的实际案例中，通过仿真实验验证了所提方法的可行性和有效性。
[Abstract]:As the basic unit of realizing the automation of industrial process, electromechanical system is the most widely used subsystem. With the progress of science and technology and the improvement of modernization level, a large number of electromechanical system structures used in industrial field are becoming more and more complex. The degree of automation is becoming higher and higher. Because of the environment and working conditions, the possibility of fault is becoming more and more, so the research on fault diagnosis of electromechanical system becomes more and more important. Taking the electromechanical system as the engineering background and the typical electromechanical system such as wind power generation system and robot as the concrete object, this paper starts from several aspects, such as improving the robustness of modeling error or disturbance, improving the sensitivity of fault diagnosis and so on. In this paper, the nonlinear fault diagnosis of electromechanical system is studied. The main work and conclusions are as follows:. Aiming at the situation that the power factor may not be 1 in the no-load grid-connection process of doubly-fed wind power generation system, a mathematical model of no-load grid-connected doubly-fed wind power generation system is established, and an integral sliding mode variable structure controller is designed. Finally, the effectiveness of the controller is proved by PSIM simulation and real-time experiment. A fault diagnosis method based on sliding mode observer and Luenberger observer is designed for sensor faults in doubly-fed wind turbine. Finally, the effectiveness of the method is proved by simulation experiments. Aiming at a class of slow-varying faults of electromechanical system, the model of electromechanical system is transformed into two subsystems by coordinate transformation. Using the complementarity of sliding mode observer and Luenberger observer, the observer is designed for the two subsystems, respectively. The robust diagnosis of weak and slowly varying faults is realized. In the design, the selection conditions of key parameters are relaxed by making full use of the LMI algorithm, and it is easier to obtain the parameters. Finally, the fault diagnosis design method is applied to the actual case of single-joint manipulator, and the feasibility and effectiveness of the proposed method are verified by simulation experiments.
【学位授予单位】：湖南工业大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TH165.3;TM614

【参考文献】