基于模糊信息粒化SVM的飞机发电机故障预测  

Aeronautical Generator Fault Prognosis Based on Fuzzy Information Granulation SVM

作　　者：和麟[1] 姜南[1] 黄潇瑶[1] 马存宝[1] 

机构地区：[1]西北工业大学航空学院,陕西西安710072

出　　处：《机械与电子》2012年第12期7-10,共4页Machinery & Electronics

基　　金：国家自然科学基金资助项目(60572173)

摘　　要：为实现复杂系统的视情维修,对故障预测的需求越来越高。将信息粒化概念引入支持向量机,提出基于模糊信息粒化支持向量机的预测算法,在分析三级无刷交流发电机脉宽调制调压器故障特征的基础上,实现对导通率的区间预测。实验结果表明,预测算法的准确率较高且可行有效,具有实际应用价值。In order to realize condition based maintenance in complex system, a higher under- standing to the fault prognosis is required. The prediction algorithm of fuzzy information granula- tion SVM is introduced in this article based on both information granulation and support vector machine. According to analyzing the conduction rate of pulse width modulation voltage regulator of three- stage brushless AC generator, the prediction algorithm can be used in making an interval failure prediction of pulse width modulation voltage regulator. The results of simulation had approved the accuracy of the prediction algorithm are high, feasible, effective and valuable.

关 键 词：故障预测 模糊信息粒化 支持向量机 三级无刷交流发电机 导通率 

分 类 号：V242.2[航空宇航科学与技术—飞行器设计]