改进阴性选择算法的风机振动故障诊断方法  

Fault Diagnosis Method for Wind Turbine Vibration Based on Improved Negative Selection Algorithm

作　　者：邬春明[1] 银海燕[1] 

机构地区：[1]东北电力大学,吉林132012

出　　处：《中国机械工程》2016年第4期479-482,共4页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(61301257);吉林省科技发展计划资助项目(2013020605GX);吉林省教育厅"十二五"科学技术研究项目(吉教科合字[2015]第250号)

摘　　要：为了正确、快速地判断风电机组振动故障类型,减小其对发电效率及人身财产安全的影响,提出了一种改进型阴性选择算法。在传统的阴性选择算法中引入马氏距离进行振动数据的初步筛选,并将算法应用于风电机组振动故障的预测。研究结果表明,改进的阴性选择算法可以更为快速、准确地判断风电机组振动的故障类型,诊断正确率达到97.5%,从而提高了风电机组运行的可靠性和发电效率。In order to correctly judge wind turbine vibration fault types, reduce its impacts on pow- er generation efficiency and personal property safety, this paper proposed an improved algorithm. In the traditional negative selection algorithm Mahalanobis distance was introduced for preliminary screening of vibration data, and the prediction algorithm was applied to the wind turbine vibration faults. At the same time with the two algorithms the diagnostic accuracy of fault diagnosis was im- proved. Studies show that the improved negative selection algorithm can be more quickly and accurate- ly determine the fault type units and the diagnostic accuracy of 97.5% are achieved. And it can im- prove the wind turbine operation reliability and power efficiency.

关 键 词：马氏距离 阴性选择 风电机组 故障诊断 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] TP391[自动化与计算机技术—控制科学与工程]