基于D-S多信息融合的风电状态监测算法研究

Research on condition monitoring algorithm of wind power based on D-S multi-information fusion

为提升风电机组状态监测的自动化、智能化数据分析能力,实现风电机组运行状态智能诊断目标,在风电状态监测数据处理过程中引入了D-S证据理论,形成了多信息融合的风机智能监测系统。首先以齿轮传动模型为例研究了齿轮故障特征的提取方法;然后分析了用隶属度函数表示相应故障特征的过程;接着建立了基于D-S数据融合的智能化故障报警系统,实现了风电机组状态监测的自动化预警,并使预警准确度提升到可工程化应用水平;最后通过现场应用来检测算法的报警准确性。结果表明,经D-S多信息融合算法过滤,可大量减少单一预警指标误报率,且已预警故障经检测验证,预警正确率达100%。

In order to improve the automatic and intelligent data analysis ability of wind turbine condition monitoring and realize the goal of intelligent diagnosis of wind turbine operation state,in the process of wind power condition monitoring data processing,D-S evidence theory is introduced to form a multi information fusion wind turbine intelligent monitoring system.Firstly,taking the gear transmission model as an example,the extraction method of gear related fault features is studied,the process of using membership functions to represent corresponding fault characteristics is analyzed;Then,an intelligent fault alarm system based on D-S data fusion is established.The automatic early warning of wind turbine condition monitoring is realized,and the accuracy of early warning is improved to the level of engineering application;Finally,the alarm accuracy of the algorithm is detected through field application.The results show that the filtering by D-S multi information fusion algorithm can greatly reduce the false alarm rate of a single early warning index,and the fault reported at the same time is 100%correct after detection.