基于PCA与蚁群算法的旋转机械故障诊断方法研究

发布时间：2018-01-17 17:01

本文关键词：基于PCA与蚁群算法的旋转机械故障诊断方法研究　出处：《湖南科技大学》2012年硕士论文　论文类型：学位论文

【摘要】：旋转机械设备被广泛应用于现代工业中，这些设备一旦出现故障将会带来巨大的经济损失。随着计算机技术的高速发展，旋转机械智能化、集成化程度越来越高，其出现的故障种类和形式越来越多，诊断时需要考虑的故障特征信息量也越来越大。由于大型复杂设备中故障信息间关系复杂多样，在众多信息中提取有效信息去除冗余信息，实现准确、高效的故障诊断一直是学术界和工程界高度关注的问题。本文结合国家自然科学基金项目，利用主元分析技术实现特征信息的有效提取，基于蚁群算法实现了设备故障快速高效的聚类诊断。主要工作如下：（1）归纳了旋转机械的核心部件——转子系统的典型故障机理，，论述了旋转机械状态信号的测量方法、故障特征提取方法及故障模式识别常用的几种方法。（2）论述了主元分析的基本原理，介绍了主元分析实现旋转机械故障特征提取的相关理论，归纳了常用的主元选取方法。基于主元分析方法实现了旋转机械故障特征的提取，分析了主元分析方法在旋转机械故障特征提取中的信息遗漏的弊端，提出了一个自适应主元选取的思路，定义了故障聚类正确率因子。（3）介绍了蚁群算法寻优的基本原理。通过蚁群算法在解决旅行商问题时寻优能力突出的特点对蚁群算法做改进。利用蚁群算法较强的寻优能力，以蚁群算法解决TSP问题为原型，建立了基于蚁群算法的旋转机械故障聚类诊断方法，设计了多因素多水平的正交实验对基于蚁群算法的旋转机械故障聚类诊断方法中初始参数做了优化。利用转子故障试验台测得的数据对基于蚁群算法的旋转机械故障聚类诊断方法进行了验证，证明能达到较好的聚类效果。（4）提出了基于主元分析与蚁群算法的旋转机械故障诊断方法模型。论述了基于核函数的主元分析旋转机械特征提取方法的实现步骤，提出了自适应主元选取方法的实现过程。建立了自适应主元选取的基于主元分析与蚁群算法的旋转机械故障聚类诊断方法，给出了该方法核心算法的伪代码和程序流程图。并结合转子故障试验台测试的数据对该方法的聚类诊断效果做了验证，证实了基于PCA与蚁群算法的旋转机械故障诊断方法的有效性。
[Abstract]:Rotating machinery equipment is widely used in modern industry, once these equipment failure will bring huge economic losses. With the rapid development of computer technology, rotating machinery intelligent. The degree of integration is becoming higher and higher, the types and forms of fault appear more and more, and the amount of fault feature information should be considered more and more. Because of the complex relationship between fault information in large-scale complex equipment. Extracting effective information from a large number of information to remove redundant information to achieve accurate and efficient fault diagnosis has been a highly concerned problem in academic and engineering circles. This paper combined with the National Natural Science Foundation project. The effective extraction of feature information is realized by principal component analysis (PCA), and the fast and efficient clustering diagnosis of equipment fault is realized based on ant colony algorithm. The main work is as follows: 1) the typical fault mechanism of rotor system, which is the core component of rotating machinery, is summarized, and the measuring method of state signal of rotating machinery is discussed. Fault feature extraction method and several common methods of fault pattern recognition. 2) the basic principle of principal component analysis (PCA) is discussed, and the related theory of fault feature extraction for rotating machinery is introduced. Based on principal component analysis (PCA), the fault features of rotating machinery are extracted, and the disadvantages of PCA in fault feature extraction of rotating machinery are analyzed. An adaptive principal component selection method is proposed, and the fault clustering accuracy factor is defined. (3) the basic principle of ant colony algorithm is introduced. The ant colony algorithm is improved by its outstanding ability in solving traveling salesman problem. Based on ant colony algorithm (ACA) to solve the TSP problem, a rotating machinery fault cluster diagnosis method based on ant colony algorithm (ACA) is established. A multi-factor and multi-level orthogonal experiment was designed to optimize the initial parameters in the fault cluster diagnosis method of rotating machinery based on ant colony algorithm. The rotating machinery based on ant colony algorithm was optimized by using the data obtained from the rotor fault test rig. The method of fault clustering diagnosis is verified. It is proved that better clustering effect can be achieved. (4) the fault diagnosis method model of rotating machinery based on principal component analysis and ant colony algorithm is proposed, and the steps of feature extraction of rotating machinery based on kernel function are discussed. The realization process of adaptive principal component selection method is put forward, and a fault cluster diagnosis method for rotating machinery based on principal component analysis and ant colony algorithm is established. The pseudo code and program flow chart of the core algorithm of this method are given, and the clustering diagnosis effect of this method is verified with the test data of rotor fault test rig. The validity of the fault diagnosis method of rotating machinery based on PCA and ant colony algorithm is proved.
【学位授予单位】：湖南科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH165.3

【参考文献】