基于模糊识别的风电双馈异步电机故障诊断方法的研究
发布时间：2018-09-03 20:21
【摘要】：风电机组双馈异步电机是风电机组的重要组成部分,是风电机组实现变速恒频的重要设备,双馈异步电机的正常运行关系着机组的运行安全。由于风电场环境恶劣、工况多变,导致发电机故障频发,而风电场一般位置偏远,设备一旦损坏,备件和更换设备的周期长。如果能够及早的发现发电机的故障,诊断出具体故障模式,并及时的调整运行模式和施行维修措施,能够缩短维修时间,降低维修费用。国内外已经针对风电机组双馈异步电机的状态监测和故障诊断进行了大量的研究工作,并且随着研究工作的向前推进,有些研究成果已经产品化,并运用到风电场中,由于研究不够深入,功能不够完善,常有误报,漏报故障的事件发生,主要原因是现有的系统主要是进行设定阈值报警,观看数据趋势变化,并没有将监测参数和工况联系起来,阈值的设定不够合理。而且这些系统给出的诊断结果只是一个初步的诊断,没有定位到具体的故障模式,不利于后续的维修计划的实施。本文在对模糊识别理论研究的基础上,提出了一套适用于风电双馈异步电机的多元模糊故障诊断流程。在流程的指导下,首先研究了双馈异步电机的结构特点和工作原理,并确定发电机典型的故障模式,包括转子不平衡、转子不对中、轴承故障、定子绕组匝间短路、转子绕组匝间短路。建立发电机系统典型故障的动力学模型,对不同的故障模式进行故障机理分析,研究故障发生时的多元故障征兆,结合故障树分析,得出故障原因和对应的维修措施,建立每种故障的故障知识库。结合风电机组变工况运行特性,提出了基于角域信号重采样和阶比分析的频域特征提取方法和基于运行工况辨识和概率分布特性的时域特征提取方法,选定与振动参数、温度参数和电气参数相关的运行参数,来划定运行区间,根据区间数据的高斯分布特性,来划定各运行区间特征参数阈值。综合考虑故障模式的多种故障征兆,结合模糊识别方法,研究融合多种故障征兆的多元模糊故障诊断方法,避免单一参数识别故障的片面性和不准确性,达到对系统故障进行准确识别的目的。识别出故障后,合理的调整运行方式和安排维修措施。在故障特征提取方法、故障模式识别方法和故障知识库的基础上,结合现场实际情况,研究开发发电机的状态监测和故障诊断系统,作为“风电机组在线状态监测和故障诊断系统TCMM V1.0"大系统的模块到现场实际应用,是理论与工程实际相结合。
[Abstract]:Wind turbine doubly-fed induction motor is an important part of wind turbine and an important equipment for wind turbine to realize variable speed and constant frequency. The normal operation of doubly-fed asynchronous motor is related to the safety of unit operation. Because the wind farm environment is bad and the working condition is changeable, the generator faults occur frequently, and the wind farm is generally located in a remote area, once the equipment is damaged, the period of spare parts and replacement equipment is long. If the generator fault can be detected early, the concrete fault mode can be diagnosed, and the operation mode and maintenance measures can be adjusted in time, the maintenance time can be shortened and the maintenance cost can be reduced. A lot of research work has been done on condition monitoring and fault diagnosis of wind turbine doubly-fed asynchronous motor at home and abroad, and with the advance of research work, some research results have been produced and applied to wind farm. Because the research is not deep enough, the function is not perfect enough, often misinformation and failure events occur, the main reason is that the existing system mainly sets the threshold alarm, watches the data trend change, The monitoring parameters are not associated with the operating conditions, and the threshold setting is not reasonable. Moreover, the diagnosis result given by these systems is only a preliminary diagnosis, which does not locate the specific fault mode, and is not conducive to the implementation of the subsequent maintenance plan. Based on the research of fuzzy recognition theory, a set of multivariate fuzzy fault diagnosis flow for wind power doubly-fed induction motor is presented in this paper. Under the guidance of the flow chart, the structural characteristics and working principle of the doubly-fed induction motor are studied firstly, and the typical fault modes of the generator are determined, including rotor imbalance, rotor misalignment, bearing fault, stator winding inter-turn short circuit, and so on. Short circuit between turns of rotor winding. The dynamic model of typical faults of generator system is established, the fault mechanism of different fault modes is analyzed, and the multiple fault symptoms when faults occur are studied. Combined with fault tree analysis, the fault causes and corresponding maintenance measures are obtained. Build the fault knowledge base for each fault. Combined with the off-condition operation characteristics of wind turbine, a frequency-domain feature extraction method based on angle domain signal resampling and order ratio analysis and a time-domain feature extraction method based on operating condition identification and probability distribution characteristics are proposed. The operating parameters related to temperature parameters and electrical parameters are used to delineate the operation interval and the threshold value of each operation interval characteristic parameter is determined according to the Gao Si distribution characteristics of the interval data. Considering the various fault symptoms of the fault mode and combining the fuzzy identification method, the multi-variable fuzzy fault diagnosis method is studied to avoid the one-sidedness and inaccuracy of single parameter fault identification. To achieve the purpose of accurate identification of system faults. After identifying the fault, adjust the operation mode and arrange the maintenance measures reasonably. On the basis of fault feature extraction method, fault pattern recognition method and fault knowledge base, combined with the actual situation in the field, the condition monitoring and fault diagnosis system of generator is researched and developed. As the module of the large-scale system of "On-line condition monitoring and fault diagnosis system TCMM V1.0" of wind turbine, it is a combination of theory and engineering practice.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TM343
本文编号：2221063
[Abstract]:Wind turbine doubly-fed induction motor is an important part of wind turbine and an important equipment for wind turbine to realize variable speed and constant frequency. The normal operation of doubly-fed asynchronous motor is related to the safety of unit operation. Because the wind farm environment is bad and the working condition is changeable, the generator faults occur frequently, and the wind farm is generally located in a remote area, once the equipment is damaged, the period of spare parts and replacement equipment is long. If the generator fault can be detected early, the concrete fault mode can be diagnosed, and the operation mode and maintenance measures can be adjusted in time, the maintenance time can be shortened and the maintenance cost can be reduced. A lot of research work has been done on condition monitoring and fault diagnosis of wind turbine doubly-fed asynchronous motor at home and abroad, and with the advance of research work, some research results have been produced and applied to wind farm. Because the research is not deep enough, the function is not perfect enough, often misinformation and failure events occur, the main reason is that the existing system mainly sets the threshold alarm, watches the data trend change, The monitoring parameters are not associated with the operating conditions, and the threshold setting is not reasonable. Moreover, the diagnosis result given by these systems is only a preliminary diagnosis, which does not locate the specific fault mode, and is not conducive to the implementation of the subsequent maintenance plan. Based on the research of fuzzy recognition theory, a set of multivariate fuzzy fault diagnosis flow for wind power doubly-fed induction motor is presented in this paper. Under the guidance of the flow chart, the structural characteristics and working principle of the doubly-fed induction motor are studied firstly, and the typical fault modes of the generator are determined, including rotor imbalance, rotor misalignment, bearing fault, stator winding inter-turn short circuit, and so on. Short circuit between turns of rotor winding. The dynamic model of typical faults of generator system is established, the fault mechanism of different fault modes is analyzed, and the multiple fault symptoms when faults occur are studied. Combined with fault tree analysis, the fault causes and corresponding maintenance measures are obtained. Build the fault knowledge base for each fault. Combined with the off-condition operation characteristics of wind turbine, a frequency-domain feature extraction method based on angle domain signal resampling and order ratio analysis and a time-domain feature extraction method based on operating condition identification and probability distribution characteristics are proposed. The operating parameters related to temperature parameters and electrical parameters are used to delineate the operation interval and the threshold value of each operation interval characteristic parameter is determined according to the Gao Si distribution characteristics of the interval data. Considering the various fault symptoms of the fault mode and combining the fuzzy identification method, the multi-variable fuzzy fault diagnosis method is studied to avoid the one-sidedness and inaccuracy of single parameter fault identification. To achieve the purpose of accurate identification of system faults. After identifying the fault, adjust the operation mode and arrange the maintenance measures reasonably. On the basis of fault feature extraction method, fault pattern recognition method and fault knowledge base, combined with the actual situation in the field, the condition monitoring and fault diagnosis system of generator is researched and developed. As the module of the large-scale system of "On-line condition monitoring and fault diagnosis system TCMM V1.0" of wind turbine, it is a combination of theory and engineering practice.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TM343
【引证文献】
相关会议论文 前2条
1 刘永扬;孙红岩;谢志江;;齿式联接不对中转子诊断理论的应用[A];设备监测与诊断技术及其应用——第十二届全国设备监测与诊断学术会议论文集[C];2005年
2 罗云林;侯学智;;基于征兆分解和模糊逻辑的飞机组件多故障诊断[A];2006中国控制与决策学术年会论文集[C];2006年
，本文编号：2221063
本文链接：https://www.wllwen.com/kejilunwen/dianlilw/2221063.html
