地铁辅助逆变器故障诊断算法研究-控制科学与工程专业论文.docx

万方数据 万方数据 摘要 随着科学技术的不断发展，工业部门出现一批生产能力强、技术先进的智能生 产设备。智能设备产业化水平已经成为评判一个国家工业现代化水准的重要标志。 但是，随着智能设备不断的自动化、智能化的同时，设备自身机构日趋复杂，设备 周围以及自身的工业环境越是复杂，原有的故障诊断技术已不能完全满足现代智能 设备对于故障诊断的高要求。本文以地铁辅助逆变器故障诊断为背景，对其可能出 现的频率变换、电压波动、脉冲暂态三种故障类型诊断研究。 本文共有五个章节内容。在第一章中主要介绍了本文选题来源、课题的背景和 研究意义。第二章节介绍了几种目前常用的信号分析方法，并客观阐述这些信号分 析方法的优缺点。第三、四、五章节为本文的核心章节，在文章中设计了地铁辅助 逆变器故障诊断仿真实验，其基本思路： 第三章节中首先，介绍了 EMD、EEMD 和 LMD 信号分析算法，并且利用这三 种信号分析算法分析地铁辅助逆变器频率变换、电压波动、脉冲暂态三种故障类型 的原始信号。然后，采用能量矩和近似熵方法提取分解信号中隐藏的故障信息并组 成故障特征向量。 第四章介绍了采用 QPSO 优化的 LSSVM 分类算法识别故障类型。第五章介绍 了采用遗传算法优化 BP 神经网络识别故障类型。在两个章节内，分别与标准 LSSVM 算法和标准 BP 神经网络进行对比实验，通过实验证明了两种优化算法相比较标准 算法能够更加有效、准确识别故障类型。 关键词：信号分析；故障诊断；QPSO-LSSVM；GABP Abstract As the development of technology, Industrial department own a batch of intelligent production equipment with strong capability and advanced technology. The degree of intelligent equipment industry has become an importance marking which measures industry modernization degree. It is an important approach to train and opening-up strategic emerging industries. However, with the device and the industrial environment more and more complicated, the methods formerly used in fault diagnosis could no longer satisfy the ever increasing demand. Hence, on the basis of the subway auxiliary inverter fault diagnosis, this paper research possible fault types such as frequency translation, voltage fluctuation and impulsive transient. This paper can be divided into five chapters. In the first chapter, the origin, engineering background and significance of the subject are introduced. The second chapter includes some common analysis methods of signal, their advantages and disadvantages were analyzed objectively. The 3rd, 4th and 5th chapters are the main content of this paper, in these chapters, the author design the fault diagnosis simulation experiment of auxiliary inverter of metro. The main thread of this paper is listed below: The 3rd chapter introduces EMD, EEM and LMD firstly. Using EMD, EEMD and LMD carry on the decomposition of or