基于灰色理论的风机变桨距驱动器故障预测

Fault prediction of variable-pitch driver in wind power generator based on grey theory

为了提高风力发电系统中的变桨距驱动器的可靠性,提出了以检测IGBT导通压降为基础,采用新息灰预测算法的变桨距驱动器故障预测方法.利用IGBT导通压降的历史数据建立GM(1,1)灰预测模型,对IGBT将来时刻的导通压降进行预测,一旦IGBT预测导通压降超过阈值,系统报警并收桨.采用等维新陈代谢算法,根据实际情况随时更新数据序列,以保证预测模型的新鲜度.提出自动变步长灰预测方法,并通过实验数据得出选取步长的经验公式.设计了IGBT导通压降检测电路,该电路抗干扰能力强、反应速度快,而且结构简单、可靠性好.实验结果表明,等维新陈代谢灰预测算法可有效预测出IGBT导通压降,提高了风机收桨的可靠性.

In order to improve the reliability of variable-pitch driver in wind power generation system,the fault prediction method based on the detection of conduction voltage drop UCE of IGBT and with adopting the innovation grey prediction algorithm was proposed.The UCE of IGBT in future time was predicted through constructing the GM（1,1） grey prediction model using the past data of UCE.When the predicted UCE value is over the threshold value,the system will offer the alarm and the turbine will shut down.To ensure the greenness degree of the prediction model,the data sequence was momentarily renewed with adopthing the equal dimension metabolism algorithm and on the basis of the practical situation.The gray prediction method with variational step-length was put forward,and the empirical formula for step-lenth selection was derived based on the experimental data.The UCE detecting circuit with the high anti-interference capability,high response speed,good reliability and simple structure was designed.The experimental result shows that the equal dimension metabolism grey prediction algorithm can effectivly predict the UCE of IGBT and improve the reliability of stopping wind power generator.