基于SCADA的风力机故障预测与健康管理技术研究

发布时间：2018-06-16 23:02

本文选题：PHM + 数据融合　；参考：《电子科技大学》2015年硕士论文

【摘要】：随着人们对环境保护日益重视,国内外风电产业呈现出高速发展态势。由于国内风电产业起步较晚,产业发展虽然快速但非常粗放,风电装备制造、风电场运营等诸多方面的产业成熟度不高,由此带来风电装备、电场的事故频发,目前已成为制约我国风电产业健康、快速发展的重要原因。因此,开展风电装备故障预测与健康管理(PHM,Prognostics and Health Management)的研究具有重要的现实意义和应用价值。以国内某风电场为对象,基于对典型风电场健康管理需求的充分的调研,提出了适合国内风力机健康管理的PHM技术研究框架。首先,按照FMECA(Failure Mode Effects and Criticality Analysis)的分析流程,完成了风力机系统的FMECA分析,给出风力机的主要部件、对应的故障模式、故障主要原因及严酷度等。其次,提出了基于风电场SCADA(Supervisory Control And Data Acquisition)数据预处理的方法。提取重要部件的状态特征之后,采用加权D_S证据理论融合技术。然后,基于模糊理论提出并建立了变权模糊综合评价模型,经过实际数据验证,评价结果贴近风力机的实际情况,验证了模型的有效性;根据风力机不同的设备特性,采用合适的诊断方法进行了故障诊断的研究,为故障预测提供支持。最后,开展了风力机故障预测的研究,基于灰色理论提出了等维灰数动态预测模型,仿真和实际应用表明,提出的预测模型提高了风力机故障预测的精度。以国内某风电场采集的实际监测数据为样本,分别对风力机的状态评价、故障诊断和状态预测进行实例验证。采用J2EE架构设计和开发了风力机故障预测与健康管理原型系统,该系统集成了实时状态评价、故障诊断和状态预测,实现了风力机的健康管理。实验结果表明,论文改进的数据融合方法、变权模糊综合评价模型及等维灰数动态预测模型等方法有效、可行,PHM原型系统很好的实现了风力机的状态评价、故障诊断和状态预测。本文的研究成果对于提高风力机运行的可靠性,降低其故障发生率,提高风电场的运营效率具有很好的实际应用价值。
[Abstract]:As people pay more and more attention to environmental protection, wind power industry at home and abroad presents a high-speed development trend. Due to the relatively late start of domestic wind power industry and the rapid but extensive industrial development, the industrial maturity of wind power equipment manufacturing, wind farm operation and many other aspects is not high, resulting in frequent accidents of wind power equipment and electric fields. At present, it has become an important reason for restricting the healthy and rapid development of wind power industry in China. Therefore, the research on wind power equipment fault prediction and health management has important practical significance and application value. Taking a domestic wind farm as an object, a PHM technology research framework suitable for domestic wind turbine health management is put forward based on the sufficient investigation on the health management requirements of typical wind farms. Firstly, according to the analysis flow of FMECAA failure Mode effects and criticality Analysis, the FMECA analysis of wind turbine system is completed, and the main components, corresponding fault modes, main causes and severity of wind turbine system are given. Secondly, a method of data preprocessing based on SCADA-SCADA supervisory control And data acquisition is proposed. After extracting the state features of important parts, the weighted DS evidence theory fusion technique is adopted. Then, based on the fuzzy theory, a fuzzy comprehensive evaluation model with variable weights is proposed and established. The evaluation results are close to the actual situation of the wind turbine, and the validity of the model is verified by the actual data, according to the different equipment characteristics of the wind turbine, the fuzzy comprehensive evaluation model is established based on the fuzzy theory. The research of fault diagnosis is carried out with proper diagnosis method, which provides support for fault prediction. Finally, the research of wind turbine fault prediction is carried out. Based on the grey theory, the dynamic prediction model of equal dimension grey number is proposed. The simulation and practical application show that the proposed prediction model improves the accuracy of wind turbine fault prediction. Taking the actual monitoring data collected from a domestic wind farm as the sample, the status evaluation, fault diagnosis and state prediction of the wind turbine are verified by examples. The prototype system of wind turbine fault prediction and health management is designed and developed with J2EE architecture. The system integrates real-time state evaluation, fault diagnosis and condition prediction, and realizes the health management of wind turbine. The experimental results show that the improved data fusion method, the variable weight fuzzy comprehensive evaluation model and the equal dimension grey number dynamic prediction model are effective, and the PHM prototype system is feasible to realize the wind turbine state evaluation. Fault diagnosis and condition prediction. The research results of this paper have good practical application value to improve the reliability of wind turbine operation, reduce its fault rate, and improve the operational efficiency of wind farm.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TM614

【参考文献】