【摘要】 分析了风电机组异常数据分布特征，设计了MKIF(Mini-batch K means-isolation forest)算法，并基于MKIF算法构建了风电机组健康劣化监测预警模型。MKIF算法将小批量K均值聚类引入搜索树的划分过程中，使用轮廓系数监督树分裂节点的数量和位置。定义了MKIF异常得分以描述数据的孤立程度，进而有效识别和剔除异常数据并基于正常运行数据建立风电机组健康基准模型。使用健康基准模型和滑动窗口算法即可对监测数据劣化程度进行评估；当窗口内劣化数据比例超过设定阈值时，触发机组健康劣化预警。以某风电机组齿轮箱油温异常导致发电性能劣化为实例，验证了模型的有效性。更多还原

【Abstract】 In this paper, the anomalous data distribution characteristics of wind turbines are analysed, the mini-batch K means-isolation forest(MKIF) algorithm is designed, and a monitoring model of wind turbine health deterioration is constructed based on the MKIF algorithm. The MKIF algorithm introduces mini batch K means clustering into the partition process of the isolation forest search tree, and uses the silhouette coefficient to supervise the number and location of the split nodes of the tree. The MKIF anomaly score is defined to describe the degree of isolation of data, which can effectively identify and eliminate abnormal data and establish a baseline model of wind turbine health based on normal operating data. Then the health baseline model and sliding window algorithm are used to evaluate the degree of deterioration of the monitoring data. When the proportion of the degraded data in the window exceeds the set threshold, the unit’s health deterioration warning is triggered. An example of a wind turbine gearbox oil temperature abnormality leading to deterioration of power generation performance verifies the effectiveness of the model in this paper.更多还原