基于改进果蝇算法优化SVM的变速箱故障诊断

Gearbox Diagnosis Based on SVM Optimized by Improved Fruit Fly Optimization Algorithm

利用果蝇算法(FOA)进行支持向量机(SVM)参数优化时容易陷入局部最优的问题,对果蝇算法进行改进设计,提出了改进果蝇算法(IFOA)。相比于FOA,IFOA增加了种群划分过程,并且不同种群按照不同步长进行位置更新,实现了算法前后期搜索能力的平衡。利用IFOA优化SVM参数并进行变速箱故障诊断,结果表明相比于FOA,IFOA诊断精度提升了6.29%;相比于其余3种改进型FOA方法,IFOA诊断精度至少提升了0.57%;相比于其它4种类型优化算法,IFOA诊断精度至少提升了2.53%,耗时缩短约39.5%。

The optimization of support vector machine(SVM)parameters using fruit fly algorithm(FOA)is easy to fall into the problem of local optimum.The improved fruit fly algorithm(IFOA)is proposed.Compared with FOA,IFOA adds a population division process,and different populations update their positions according to the asynchronous length,so as to achieve the balance of the search ability of the algorithm before and after.SVM parameters were optimized by IFOA and gearbox fault diagnosis was carried out.The results show that compared with FOA,the diagnosis accuracy of IFOA is improved by 6.29%.Compared with the other three improved FOA methods,the diagnostic accuracy of IFOA improved by at least 0.57%.Compared with the other four types of optimization algorithms,THE diagnosis accuracy of IFOA is improved by at least 2.53%and the time consumption is shortened by 39.5%.