离心式压缩机轴心轨迹自动识别及其故障诊断系统

发布时间：2018-04-26 07:10

本文选题：离心式压缩机 + 轴心轨迹　；参考：《浙江大学》2013年硕士论文

【摘要】：本文结合国家重点基础研究发展计划项目“复杂空气分离类成套装备超大型化与低能耗化的关键科学问题”(编号：2011CB706505)和国家自然科学基金项目“微弱冲击信号的提取与识别技术”(编号：51175466)对离心式压缩机轴心轨迹的提纯、识别及其故障诊断系统进行了研究。第一章：介绍了离心式压缩机的工作原理和结构特点；阐述了离心式压缩机轴心轨迹自动识别与其故障诊断系统的研究意义；分析了轴心轨迹提纯、特征提取、模式识别的现有方法；提出了论文研究内容和全文的框架。第二章：分析了几种常见故障的轴心轨迹特点,给出了转子不平衡、转子不对中、支持系统松动、动静碰摩等几种常见故障的力学模型,探讨了对应的轴心轨形状及其故障特征。第三章：分析了基于小波变换、小波包的去噪方法。探讨了谐波小波的特点,引入了一种基于汉宁窗和小波包的谐波小波改进方法,以此提纯轴心轨迹,并进行了实验分析。第四章：分析了小波矩的构造方法；提出了基于离散尺度的小波矩特征选择方法；以所选择的小波矩特征为输入、以支持向量机为模式识别工具,建立了轴心轨迹的识别方法,并进行了实验验证。第五章：给出了离心式压缩机故障诊断系统总体架构设计。基于LabVIEW设计了系统中的主要功能模块,进行了系统开发。第六章：总结全文内容,对下一步研究提出了展望。
[Abstract]:In this paper, the National key basic Research and Development Program, "key Scientific problems in the Ultra-large and low Energy consumption of complex Air Separation complete sets" (No.: 2011CB706505) and the National Natural Science Foundation Project "weak impact signal" Extraction and Identification techniques "(No. 51175466) for the purification of the axial trajectory of centrifugal compressors, Identification and fault diagnosis system are studied. In the first chapter, the working principle and structure characteristics of centrifugal compressor are introduced, the research significance of automatic identification of axis track and fault diagnosis system of centrifugal compressor is expounded, the purification of axis trajectory and feature extraction are analyzed. The existing methods of pattern recognition and the research contents and the framework of the thesis are presented. In the second chapter, the characteristics of axis track of several common faults are analyzed, and the mechanical models of rotor unbalance, rotor misalignment, support system loosening, dynamic and static impact are given. The corresponding axial rail shape and its fault characteristics are discussed. Chapter 3: wavelet packet denoising method based on wavelet transform is analyzed. This paper discusses the characteristics of harmonic wavelet and introduces an improved method of harmonic wavelet based on Hanning window and wavelet packet to purify the orbit of axis center. Chapter 4: the construction method of wavelet moment is analyzed, the wavelet moment feature selection method based on discrete scale is put forward, the selected wavelet moment feature is used as input, and the support vector machine is used as pattern recognition tool. The experimental results are verified. Chapter 5: the overall frame design of centrifugal compressor fault diagnosis system is given. The main function modules of the system are designed based on LabVIEW, and the system is developed. Chapter six: summarize the content of the full text and put forward the prospect of the next research.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TH452;TP277

【参考文献】