基于多特征融合的风电叶片实时故障诊断研究

Research on Real-time Fault Diagnosis of Wind Turbine Blades Based on Multi-feature Fusion

  • 摘要: 为提高风电机组的叶片(以下简称风电叶片)故障实时诊断准确率,降低诊断的误报率和漏报率,本文以叶片振动信号为基础,统计风电叶片故障后特征变化概率和变化量,总结得到风电叶片损伤后变化明显的特征。为消除单一特征对诊断结果的绝对判断,降低变化不明显的特征影响力,本文提出一种多特征融合计算方法,根据融合特征对风电叶片进行实时故障诊断。研究结果及应用表明,采用多特征融合计算方法可以较好反映叶片出现叶尖开裂、叶片裂纹等故障,对风电叶片的实时故障诊断具有参考价值。

     

    Abstract: To improve the real-time diagnostic accuracy of wind turbine blades (hereinafter referred to as wind turbine blades) and reduce the false alarm and false alarm rates in diagnosis, based on the blade vibration signal, the probability and amount of characteristic changes after wind turbine blade failure are statistically analyzed, and the obvious characteristics of changes after wind turbine blade damage are summarized. To eliminate the absolute judgment of single features on diagnostic results and reduce the influence of inconspicuous changes in features, a multi feature fusion calculation method is proposed to perform real-time fault diagnosis of wind turbine blades based on fused features. The research results and applications indicate that the use of multi-feature fusion calculation method can effectively reflect faults such as blade tip cracking and blade cracks, and has reference value for real-time fault diagnosis of wind turbine blades.

     

/

返回文章
返回