| 引用本文: | 郝勇,吴文辉,商庆园.基于变分模态分解和支持向量机的滚动轴承品质评估[J].控制理论与应用,2020,37(7):1544~1551.[点击复制] |
| HAO Yong,WU Wen-hui,SHANG Qing-yuan.Rolling bearing quality evaluation based on variational mode decomposition and support vector machines[J].Control Theory and Technology,2020,37(7):1544~1551.[点击复制] |
|
| 基于变分模态分解和支持向量机的滚动轴承品质评估 |
| Rolling bearing quality evaluation based on variational mode decomposition and support vector machines |
| 摘要点击 1662 全文点击 520 投稿时间：2019-07-21 修订日期：2019-12-19 |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| DOI编号 10.7641/CTA.2020.90594 |
| 2020,37(7):1544-1551 |
| 中文关键词 滚动轴承 变分模态分解 排列熵 时域特征 支持向量机 |
| 英文关键词 rolling bearing variational mode decomposition permutation entropy time domain characteristics support vector machines |
| 基金项目 国家自然科学基金项目(21265006, 51665013)资助. |
|
| 中文摘要 |
| 针对滚动轴承品质评估过程中振动信号代表性特征提取不充分且模式识别方法精度低等不足, 提出了基
于变分模态分解(VMD)和支持向量机(SVM)的滚动轴承品质评估方法, 首先, 对3个品质等级的轴承样品进行振动
信号的采集; 其次, 计算滚动轴承振动信号的有效值、峰值和峭度值3个时域指标(TDI), 并采用VMD方法将信号分
解为4个有限带宽模态函数(BIMF), 并分别计算其排列熵(PE)值; 最后, 将3个时域指标和4个排列熵值共计7个特征
作为SVM的输入变量构建轴承品质等级预测评估模型. 实验结果表明: 与TDI–PE–SVM模型相比, TDI–VMD–PE–
SVM轴承品质评估模型更优, 识别率由83.33%提高到93.33%, VMD方法有效地提高了振动信号的分辨率, 有利于
轴承振动信号细节特征信息的提取. |
| 英文摘要 |
| Aiming at the insufficient extraction of representative features of vibration signals and the low accuracy of
pattern recognition methods in the process of quality evaluation of rolling bearings, a method for evaluating the quality
of rolling bearings based on variational mode decomposition (VMD) and support vector machines (SVM) was proposed.
Firstly, vibration signals of the bearing samples with three quality grades were collected. Secondly, the time domain
indexs (TDI) that effective value, peak value and kurtosis value were calculated, and vibration signals were decomposed by
VMD. Each vibration signal was decomposed into four band-limited instrinsic mode function (BIMF), and the permutation
entropy (PE) was calculated respectively. Finally, a total of 7 features including three time domain characteristics and four
PE features were used as inputs of SVM to establish models for quality level prediction and evaluation. The experimental
results show that compared with the TDI–PE–SVM model, the TDI–VMD–PE–SVM bearing quality evaluation model is
better, and the recognition rate is improved from 83.33% to 93.33%. The VMD method effectively improves the resolution
of vibration signals, which is conducive to the extraction of detailed characteristic information of bearing vibration signals. |
|
|
|
|
|