摘要
基于经验模态分解(EMD)算法的递归特性提出优化变分模态分解(VMD)算法,结合能量熵方法构建多模态特征矩阵,通过鲸鱼算法优化的支持向量机技术(OSVM)实现轴承的故障诊断,并验证所提算法的有效性。结果表明:基于VMD算法和能量熵构建的多模态特征矩阵对故障的区分度优于EMD算法和能量熵方法;与现有方法相比,所提VMD-OSVM算法在变负载和噪声环境下的诊断准确率分别高出13.8%与30%,体现了该算法良好的鲁棒性和泛化性能;在相同计算资源下,所提VMD-OSVM算法的运行时间更短,效率更高。
Based on the recursive characteristics of empirical mode decomposition(EMD)algorithm,the optimized variational mode decomposition(VMD)algorithm was proposed,so as to establish the multi-modal characteristic matrix combining with the energy entropy method.Finally,the fault diagnosis of rolling bearings was realized using support vector machine technology(OSVM)optimized by the whale algorithm,during which the effectiveness of the proposed algorithm was verified.Results show that the multi-modal characteristic matrix based on VMD algorithm and energy entropy method is superior to the EMD algorithm and energy entropy method in fault discrimination.Compared with the existing methods,accuracy of the proposed VMD-OSVM algorithm is 13.8%and 30%higher under variable load and noise environments,respectively,which reflects the good robustness and generalization performance of the algorithm.Under the same computing resources,the proposed method saves running time,and is more efficient.
作者
金江涛
许子非
李春
缪维跑
JIN Jiangtao;XU Zifei;LI Chun;MIAO Weipao(School of Energy and Power Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China;Shanghai Key Laboratory of Multiphase Flow and Heat Transfer in Power Engineering,Shanghai 200093,China)
出处
《动力工程学报》
CAS
CSCD
北大核心
2021年第3期214-220,243,共8页
Journal of Chinese Society of Power Engineering
基金
国家自然科学基金资助项目(51976131,51676131)
上海市“科技创新行动计划”地方院校能力建设资助项目(19060502200)。
