基于统计区间划分的风电机组齿轮箱轴承预警方法研究

Early Warning Method of Wind Turbines gear box bearing Based on Statistical Division

针对风电机组运行工况复杂、单机容量小、状态监测设备安装成本高的特点，本史通过SCADA数据实现了齿轮箱轴承故障预警。提出了基于统计区间划分方法，通过环境温度和有功功率进行二维区间划分来预测齿轮箱轴承温度，利用滑动窗口统计异常比率作为触发援信度水平汁箅的指标，引人不同同定时段的平均值构建滑动窗口对比各自预警结果，克服了单一阈值预警时的误报和漏报问题。最后，通过实际案例对算法进行了验证，得出结论，该方法可通过机组历史SCADA数据，提取数据特性实时监测齿轮箱轴承状态。

In view of the characteristics of complex operating conditions, small single capacity and high installation cost of condition monitoring equipment, this paper realizes the fault warning of gear box bearing by SCADA data. The method based on the statistical interval partitioning is proposed, The bearing temperature of the gearbox is predicted by two-dimensional interval partitioning by the ambient temperature and active power, and the statistical anomaly ratio of sliding window is used as the index of triggering confidence level, The average value of different fixed periods is introduced to compare the respective early warning results. The problem of false positives and false negatives in single threshold warning is overcome. Finally, the algorithm is validated by the actual case, and the conclusion is drawn that the method can be used to monitor the bearing status of the gearbox by extracting the data from the historical SCADA data of the unit.