摘要
风电机组变桨系统是风电机组发生故障最频繁的部件之一,对其故障类型的精确诊断能够提高风电机组维护计划的效率。针对异步电机和行星齿轮箱的各种故障类型,提出了一项以风电机组三相电流数据为基础的多分量故障诊断方法。该方法通过深度自动编码器从三相电流数据中提取特征向量,并采用支持向量机进行故障分类。上述方法以风电机组变桨驱动器为例进行验证,实验结果表明在变负载和变转速环境下,上述方法能够实现对风电机组变桨系统故障类型的准确识别和诊断。
The rotor system of wind turbine is the most fragile parts of wind turbine, and accurate diagnosis of its fault categories can improve the effectiveness of wind turbine maintenance. A multi-component fault diagnosis method based on the three-phase currents data of wind turbine was proposed for various fault categories of induction motor and planetary gearbox. The method extracted feature vectors from three-phase currents data by deep auto-encoder and used support vector machine for fault classification. The above method was verified by an example of a rotor driver of wind turbine. The experimental results show that the method can accurately identify and diagnose the fault categories of rotor driver under variable load and variable speed environment.
作者
张真真
吴立东
陈晓敏
徐志轩
曹善桥
ZHANG Zhenzhen;WU Lidong;CHEN Xiaomin;XU Zhixuan;CAO Shanqiao(China Datang Corporation Renewable Energy Science and Technology Research Institute Co.,Ltd.,Xicheng District,Beijing 100052,China)
出处
《分布式能源》
2021年第3期70-75,共6页
Distributed Energy
基金
中国大唐集团新能源科学技术研究院有限公司科技项目(新能源监控与大数据中心多源融合远程专家诊断系统研发)。
关键词
变桨系统
三相电流
深度自动编码器
支持向量机
pitch system
three-phase current
deep auto-encoder
support vector machine
