基于多类CS-SVM直驱风力发电机轴承故障诊断研究
发布时间：2019-05-24 22:05
【摘要】：随着风电产业的快速发展,风力发电技术已经成为国内外的研究热点。由于运行中的风电机组事故率高,迫切需要一种高效的机组故障诊断技术,提高机组可靠性、运行效率,降低维护费用、停机时间。这对风电产业的发展具有深远的影响。论文介绍了故障诊断的基本原理,总结了直驱风力发电机常见故障及故障机理,分析现有特征提取的基本原理,在此基础上提出改进小波包特征提取法,该方法利用信号的频谱分析,先确定分析步长和分析位置,再进行小波包分解。通过小波包特征提取、多源特征提取和改进小波包特征提取的对比分析,证明了改进特征提取方法的有效性。分析标准支持向量机、多类支持向量机,针对直驱风力发电机轴承样本类别分布不平衡的问题,结合风力发电故障诊断的研究发展方向,提出一种基于多类代价敏感支持向量的故障诊断方法。分析网格寻优算法、粒子群寻优算法和遗传寻优算法,在此基础上,采用改进的粒子群寻优算法进行代价敏感支持向量机的三参数寻优。经过与传统寻优算法的对比分析,证明了改进算法寻优速度更快。在前面内容的基础上,构建直驱风力发电机轴承故障诊断模型,通过对其故障样本集的模拟,分析模型的故障敏感性、鲁棒性、新增类型样本的识别能力,证明了多类代价敏感支持向量机故障诊断模型的优异性能。总结了直驱风力发电机轴承故障诊断方法有待于完善和进一步研究的问题。论文所做的工作对直驱风力发电机轴承故障诊断具有重要的参考价值。
[Abstract]:With the rapid development of wind power industry, wind power generation technology has become a hot research topic at home and abroad. Because of the high accident rate of wind turbine in operation, it is urgent to need an efficient fault diagnosis technology to improve the reliability, operation efficiency, maintenance cost and downtime of the unit. This has a profound impact on the development of wind power industry. This paper introduces the basic principle of fault diagnosis, summarizes the common faults and fault mechanism of direct drive wind turbine, analyzes the existing basic principles of feature extraction, and puts forward an improved wavelet packet feature extraction method. In this method, the analysis step size and analysis position are determined by using the spectrum analysis of the signal, and then the wavelet packet decomposition is carried out. Through the comparative analysis of wavelet packet feature extraction, multi-source feature extraction and improved wavelet packet feature extraction, the effectiveness of the improved feature extraction method is proved. The standard support vector machine (SVM) and multi-class support vector machine (SVM) are analyzed to solve the problem of unbalanced distribution of bearing samples for direct drive wind turbines, combined with the research and development direction of wind power fault diagnosis. A fault diagnosis method based on multi-class cost-sensitive support vectors is proposed. The grid optimization algorithm, particle swarm optimization algorithm and genetic optimization algorithm are analyzed. on this basis, the improved particle swarm optimization algorithm is used to optimize the three parameters of cost sensitive support vector machine. Compared with the traditional optimization algorithm, it is proved that the improved algorithm is faster. On the basis of the previous contents, the bearing fault diagnosis model of direct drive wind turbine is constructed. through the simulation of its fault sample set, the fault sensitivity, robustness and recognition ability of the new type samples are analyzed. The excellent performance of multi-class cost-sensitive support vector machine fault diagnosis model is proved. This paper summarizes the problems that need to be improved and further studied in the bearing fault diagnosis method of direct drive wind turbine. The work done in this paper has important reference value for bearing fault diagnosis of direct drive wind turbine.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM315
，
本文编号：2485222
[Abstract]:With the rapid development of wind power industry, wind power generation technology has become a hot research topic at home and abroad. Because of the high accident rate of wind turbine in operation, it is urgent to need an efficient fault diagnosis technology to improve the reliability, operation efficiency, maintenance cost and downtime of the unit. This has a profound impact on the development of wind power industry. This paper introduces the basic principle of fault diagnosis, summarizes the common faults and fault mechanism of direct drive wind turbine, analyzes the existing basic principles of feature extraction, and puts forward an improved wavelet packet feature extraction method. In this method, the analysis step size and analysis position are determined by using the spectrum analysis of the signal, and then the wavelet packet decomposition is carried out. Through the comparative analysis of wavelet packet feature extraction, multi-source feature extraction and improved wavelet packet feature extraction, the effectiveness of the improved feature extraction method is proved. The standard support vector machine (SVM) and multi-class support vector machine (SVM) are analyzed to solve the problem of unbalanced distribution of bearing samples for direct drive wind turbines, combined with the research and development direction of wind power fault diagnosis. A fault diagnosis method based on multi-class cost-sensitive support vectors is proposed. The grid optimization algorithm, particle swarm optimization algorithm and genetic optimization algorithm are analyzed. on this basis, the improved particle swarm optimization algorithm is used to optimize the three parameters of cost sensitive support vector machine. Compared with the traditional optimization algorithm, it is proved that the improved algorithm is faster. On the basis of the previous contents, the bearing fault diagnosis model of direct drive wind turbine is constructed. through the simulation of its fault sample set, the fault sensitivity, robustness and recognition ability of the new type samples are analyzed. The excellent performance of multi-class cost-sensitive support vector machine fault diagnosis model is proved. This paper summarizes the problems that need to be improved and further studied in the bearing fault diagnosis method of direct drive wind turbine. The work done in this paper has important reference value for bearing fault diagnosis of direct drive wind turbine.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM315
，
本文编号：2485222
本文链接：https://www.wllwen.com/kejilunwen/dianlilw/2485222.html
教材专著
