摘要
为了提升滚动轴承的识别效率,保障机械设备的正常安全运行,文中提出了一种基于鲸鱼优化算法与多核学习的支持向量机相结合的故障诊断方法。在研究过程中,由于滚动轴承信号的强非线性与故障为多分类问题,文中引入了多核学习方法,通过将多核函数映射到高维空间,构建由多个特征空间向量组合成的新的空间。新的空间充分利用各子空间的映射能力,对于故障分类这一复杂问题有了较强的适应性,弥补了单核函数非线性差的缺陷。为了解决多核参数选择问题,引入了鲸鱼优化算法,提升了模型的训练效率。在6205-2RS型深沟球轴承实验平台上,通过与单核SVM、GS-MKSVM等算法的对照实验,验证了该算法具有较好的分类识别能力,算法的精度达到了94.4%,相较于其他算法提升了约15%。
In order to improve the recognition efficiency of rolling bearing and ensure the normal and safe operation of mechanical equipment,this paper proposes a fault diagnosis method based on whale optimization algorithm and multi-core learning support vector machine.In the research process,because of the strong nonlinearity of rolling bearing signal and the multi classification problem of fault,this paper introduces the multi-core learning method.By mapping the multi-core function to the high-dimensional space,a new space composed of multiple eigenspace vectors is constructed.The new space makes full use of the mapping ability of each subspace,and has a stronger adaptability for the complex problem of fault classification.It makes up for the defect of nonlinear difference of single kernel function.In order to solve the multi-core parameter selection problem,whale optimization algorithm is introduced to improve the training efficiency of the model.On the 6205-2 RS deep groove ball bearing experimental platform,through the contrast experiment with single core SVM,GS-MKSVM and other algorithms,it is verified that the algorithm in this paper has better classification and recognition ability.The accuracy of the algorithm reaches 94.4%,which is about 15%higher than other algorithms.
作者
王建国
陈锴
张文兴
秦波
WANG Jianguo;CHEN Kai;ZHANG Wenxing;QIN Bo(Inner Mongolia University of Science and Technology,Baotou 014010,China)
出处
《电子设计工程》
2021年第3期31-35,共5页
Electronic Design Engineering
基金
国家自然科学基金地区项目(51865045)。
