一种基于K均值聚类分析的风机退化模态识别

A K-Means Clustering Analysis of Wind Turbine Degenerate State Recognition

为提高风机预警维护的及时性与准确性,需对风机运营状态、退化模式进行有效识别,提出一种基于K均值聚类分析的风机退化模态识别。基于统计识别理论,引入K均值聚类分析对历史数据与状态信息学习分类,通过调整类内紧密度识别齿轮箱退化状态,利用风机模拟平台进行退化状态评估实验,划分退化区间,验证齿轮箱性能评估的有效性。实验结果表明基于K均值聚类分析的风机退化识别模型可有效识别风机运营模态,划分退化区域,为建立风机维护模型提供更为精确的科学依据。

In order to improve timeliness and accuracy of wind turbine early warning and maintenance,it is necessary to identify the operating state and degenerate modal. A degenerate modal identification based on K-means clustering has been proposed. Based on statistical recognition,K-means clustering analysis is used to classify historical data and state information.Identify the degenerate state of the gear box by adjusting the tight density in classes. The degenerate state evaluation experiment is carried out by wind turbine simulation platform,and the degenerate interval is divided to verify the effectiveness of the gear box performance evaluation. The results show that wind turbine degradation recognition model based on K-mean clustering analysis can effectively identify operation mode and divide the degraded area,and also provide more precise scientific basis for establishing wind turbine maintenance model.