一种风机变桨系统运行状态异常识别方法

摘要

针对风电场数据量大、类型多且数据结构复杂的特点,采用分段多项式曲线拟合的方法,建立风电机组变桨系统的健康模型.依据健康模型,对实时工况的参数变化趋势进行预测,然后采用滑动窗口残差均值分析法,对实测值和预测值之间的残差进行阈值判断,并结合现场经验对变桨系统的运行状态进行异常识别.历史数据整理分析和Matlab仿真平台验证表明:该方法对于识别风机的异常运行状态具有可行性.%Due to the complex structure,large volume and various types of the wind farm data,the health model for variable pitch system of wind turbine generators system is built by using piecewise polynomial curve fitting. Based on the health model,the varying trend of parameters about real time operating conditions is predicted; then by using the sliding window residual mean analysis method,the threshold judgment for residual error between actual value and predicted value is conducted; and combining the practical experience in the field, the abnormal operation state of the variable pitch system is identified. The analysis of historical data and verification of Matlab simulation platform show that the result shows that the method is feasible to identify the abnormal running state of wind turbine.