基于样本熵的风力发电机早期故障检测

针对发电机定子匝间短路和转子断条等早期故障特征具有幅值小、非稳态、易受工况影响等特点,引入样本熵算法实现风力发电机定子电流和电磁转矩信号特征提取,并模拟不同负载条件下故障信号,实现定量参数分析。分析结果表明,样本熵算法适用于在变工况及噪声干扰条件下,对短数据参量进行分析并实现故障特征定量描述,可用于风力发电机早期故障检测和实时在线监测。

Incipient Fault Detection of Generators in Wind Turbines Based on Sample Entropy

In order to test the generator incipient faults,such as stator windings inter-turn and broken rotor bars,which have such characteristics with small amplitude,non-stationary and susceptible to load variations,a sample entropy algorithm is introduced to perform fault feature extraction from the stator current and electromagnetic torque signals of generators in wind turbines.The proposed method is used to analyze fault signals under different load conditions to realize the quantitative representation of faulty signals.The results demonstrate that sample entropy algorithm is applicable to achieve fault characteristics quantification with data of short length,especially under varying operations and noise environments and it has great potentials in early fault detection and real-time online monitoring.