基于最小二乘法的风机变桨系统故障预警

Fault Early Warning of Wind Turbine Pitch System Based on Least Square Method

提出了一种新的风机变桨系统故障预警方法.通过最小二乘法拟合历史数据曲线,形成正常运行状态下的健康模型.采用基于欧氏距离的方法对输出残差进行统计分析,确定预警阈值并计算异常率作为触发预警指标.最后,以风速和1#叶片桨距角为例,选择处于额定风速到切出风速之间的数据作为研究对象,经过在Matlab仿真表明:该方法能够准确地对风机变桨系统进行故障预警,具有一定的可行性.

The paper puts forward a new fault warning method of the draught fan variable propeller system by the least squares method fitting historical data curve so it can become a health model under the normal operation.Then according to Euclidean Distance method,the paper satistically analyzes the output residual error,determines the warning threshold value and takes the abnormal rate calculation as trigger warning index. Finally taking the wind speed and 1# blade pitch angle as an example,the paper chooses the data between rated wind speed and cut out wind speed as the research object,through the Matlab simulation test,which shows that this method can accurately warn the fault of draught fan variable propeller system forecast and feasibiility.