基于GWO优化的CS-SVM轴承故障诊断
|
摘要
灰狼算法(GWO)作为新型寻优算法,可用于轴承故障诊断。提出了采用GWO优化代价敏感支持向量机(CS-SVM)的诊断模型。通过经验模态分解(EMD)及主成分分析(PCA)进行特征提取并实现特征的降维,GWO优化CS-SVM参数来提升故障分类的准确率。以西储大学轴承数据为例,将比例为4∶1的训练样本和测试样本带入GWO优化的CS-SVM模型,诊断测试的准确率为96.67%,相比于传统PSO算法的准确率有所提升,收敛速度更快,表明了GWO优化的CS-SVM具有优越性。由此可以得出,GWO可用于轴承故障诊断的研究,验证了该算法模型的有效性。
Grey wolf optimizer(GWO), as a new optimization algorithm, can be used for bearing fault diagnosis. A diagnostic model was illustrated for optimize cost-sensitive support vector machine(CSSVM) by GWO. Empirical mode decomposition(EMD) and principal component analysis(PCA) were used to extract features and reduce dimensionality,GWO optimized CS-SVM parameters to improve the accuracy of fault classification. Taking bearing data of Case Western Reserve University as an example,take the training set and the test set with a ratio of 4∶1 into the model. The correctness of the signal is96.67%. Compared with the traditional PSO algorithm, the accuracy is improved and the convergence speed is faster, it shows the superiority of CS-SVM optimized by GWO. It can be concluded that the grey wolf optimizer can be used in bearing fault diagnosis research, which verifies the effectiveness of the proposed algorithm model.
Grey wolf optimizer(GWO), as a new optimization algorithm, can be used for bearing fault diagnosis. A diagnostic model was illustrated for optimize cost-sensitive support vector machine(CSSVM) by GWO. Empirical mode decomposition(EMD) and principal component analysis(PCA) were used to extract features and reduce dimensionality,GWO optimized CS-SVM parameters to improve the accuracy of fault classification. Taking bearing data of Case Western Reserve University as an example,take the training set and the test set with a ratio of 4∶1 into the model. The correctness of the signal is96.67%. Compared with the traditional PSO algorithm, the accuracy is improved and the convergence speed is faster, it shows the superiority of CS-SVM optimized by GWO. It can be concluded that the grey wolf optimizer can be used in bearing fault diagnosis research, which verifies the effectiveness of the proposed algorithm model.
引文
[1]王振华.基于ESMD和SVM的滚动轴承故障诊断研究[D].汉中:陕西理工大学,2018.
[2]段礼祥,谢骏遥,王凯,等.基于不同工况下辅助数据集的齿轮箱故障诊断[J].振动与冲击,2017,36(10):104-108+116.
[3]万建武,杨明,陈银娟.代价敏感的半监督Laplacian支持向量机[J].电子学报,2012,40(7):1 410-1 415.
[4]唐明珠,阳春华,桂卫华.基于改进的QBC和CS-SVM的故障检测[J].控制与决策,2012,27(10):1 489-1 493.
[5]何大伟,彭靖波,胡金海,等.基于改进FOA优化的CS-SVM轴承故障诊断研究[J].振动与冲击,2018,37(18):108-114.
[6]姜天华.混合灰狼优化算法求解柔性作业车间调度问题[J].控制与决策,2018,33(3):503-508.
[7]左剑,张程稳,肖逸,等.基于灰狼优化算法的多机电力系统稳定器参数最优设计[J].电网技术,2017,41(9):2 987-2 995.
[8]李洪燕,陈步英,赵美枝.数据挖掘技术在煤矿机械故障诊断中的应用[J].煤矿机械,2012,33(12):259-261.
[9]赵伟.基于IFOA算法的SVM参数优化及其应用[J].计算机系统应用,2015,24(6):207-210.
[2]段礼祥,谢骏遥,王凯,等.基于不同工况下辅助数据集的齿轮箱故障诊断[J].振动与冲击,2017,36(10):104-108+116.
[3]万建武,杨明,陈银娟.代价敏感的半监督Laplacian支持向量机[J].电子学报,2012,40(7):1 410-1 415.
[4]唐明珠,阳春华,桂卫华.基于改进的QBC和CS-SVM的故障检测[J].控制与决策,2012,27(10):1 489-1 493.
[5]何大伟,彭靖波,胡金海,等.基于改进FOA优化的CS-SVM轴承故障诊断研究[J].振动与冲击,2018,37(18):108-114.
[6]姜天华.混合灰狼优化算法求解柔性作业车间调度问题[J].控制与决策,2018,33(3):503-508.
[7]左剑,张程稳,肖逸,等.基于灰狼优化算法的多机电力系统稳定器参数最优设计[J].电网技术,2017,41(9):2 987-2 995.
[8]李洪燕,陈步英,赵美枝.数据挖掘技术在煤矿机械故障诊断中的应用[J].煤矿机械,2012,33(12):259-261.
[9]赵伟.基于IFOA算法的SVM参数优化及其应用[J].计算机系统应用,2015,24(6):207-210.