基于粒子群优化SVM的飞机发电机故障诊断  

Aeronautical Generator Fault diagnosis Based on Particle Swarm Optimization SVM

作　　者：和麟[1] 梁丽嫒 黄潇瑶[2] 马存宝[1] 

机构地区：[1]西北工业大学航空学院,西安710072 [2]西安微电子技术研究所,西安710054

出　　处：《计算机测量与控制》2013年第12期3237-3239,共3页Computer Measurement &Control

摘　　要：飞机交流发电机是飞机电源系统中最重要的部件,因此对飞机发电机的故障诊断至关重要;针对支持向量机(SVM)故障诊断中存在的参数寻优随意性、盲目性和效率低的问题,将粒子群优化算法引入SVM参数寻优,深入研究了基于粒子群优化SVM的飞机发电机故障诊断算法;以主发电机的阻尼绕组电流和励磁绕组电流为特征量,提取故障特征样本,通过分析三级无刷交流发电机的交流励磁机励磁绕组故障,实现对飞机发电机的故障诊断;算例结果表明,该诊断算法能有效诊断出故障,并且有较高的准确率和诊断效率。Aircraft AC generator is the most important component of aircraft electrical power system, so fault diagnosis of the AC gener- ator is vital. To solve the problems of the randomness, blindness and inefficiency in the parameter optimization process of the supprot vector machine （SVM） algorithm, this paper introduces the particle swarm optimization （PSO） into parameter optimization to improve the SVM algorithm, which is then utilized for the fault diagnosis of aircraft AC generator. The fault feature samples are extracted with the characteris- tic parameters regarding the damping and exciting windings currents. Then The fault diagnosis of the aircraft AC generator is conducted by means of analyzing the fault mode of exciting windings of the AC exciter in the three--stage brushless AC generator. Illustrative example shows that this approach is capable to daignose the fault mode, while has a satisfactory accuracy and efficiency.

关 键 词：故障诊断 支持向量机 粒子群优化 三级无刷交流发电机 励磁绕组故障 

分 类 号：V242.2[航空宇航科学与技术—飞行器设计]