针对采用传统特征指标进行故障诊断准确率较低的问题，提出了一种基于混合标度律特征和改进支持向量机的滚动轴承智能故障诊断方法。首先，利用超阶分析得到指示故障的标度律指标，并将其与常规特征指标相结合构造混合特征指标矩阵，提升特征指标对故障的区分度。其次，采用支持向量机（Support Vector Machines, SVM）对构造的混合特征矩阵进行分类，利用粒子群优化算法对SVM中重要参数进行优化。最后，利用滚动轴承试验台对提出的滚动轴承智能故障诊断方法进行验证。结果表明：与常规特征相比，利用构造混合特征指标得到的训练准确率提高了13%，测试准确率提高了23%。所提方法不仅能识别不同故障类型，而且能对同一故障不同损伤程度进行识别，有望进一步实现滚动轴承故障定量诊断。
关键词：智能故障诊断；超阶分析；混合特征指标；粒子群优化；支持向量机
Abstract
To address the problem of low accuracy of fault diagnosis using traditional feature indexes, a new intelligent fault diagnosis method of rolling bearing is proposed based on hybrid scale exponent index and improved support vector machine. First, the scale exponent index for indicating fault is obtained by using the super order analysis, and the hybrid characteristic index matrix is constructed by combining it with the conventional characteristic indexes, so as to improve the discrimination of the characteristic index to the fault. Second, Support Vector Machine (SVM) is used to classify the constructed mixed vectors, and particle swarm optimization algorithm is used to optimize the important parameters of SVM. Finally, the proposed intelligent fault diagnosis method of rolling bearing is verified by using the rolling bearing test bench. The results show that the training accuracy and testing accuracy using the hybrid feature indexes are improved by 13% and 23%, respectively, compared with the conventional feature indexes. The proposed method can not only identify the fault types, but also identify the damage degree of the same fault, which emerges to further realize the quantitative fault diagnosis of rolling bearing.
Key words: Intelligent fault diagnosis; Super-order analysis; Hybrid characteristic indexes; Particle swarm optimization; Support vector machine
关键词
智能故障诊断 /
超阶分析 /
混合特征指标 /
粒子群优化 /
支持向量机
{{custom_keyword}} /
Key words
Intelligent fault diagnosis /
Super-order analysis /
Hybrid characteristic indexes /
Particle swarm optimization /
Support vector machine
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 刘志亮，潘登，左明健，等. 轨道车辆故障诊断研究进展[J]. 机械工程学报，2016，52(14)：134-146.
Liu Zhi-liang, Pan Deng, Zuo Ming-jian, et al. A review on fault diagnosis for rail vehicles[J]. Journal of Mechanical Engineering, 2016, 52(14): 134-146.
[2] 张立国，李盼，李梅梅，等. 基于ITD模糊熵和GG聚类的滚动轴承故障诊断[J]. 仪器仪表学报，2014，35(11)：2624-2632.
Zhang Li-guo, Li Pan, Li Mei-mei, et al. Fault diagnosis of rolling bearing based on ITD fuzzy entropy and GG clustering[J]. Chinese Journal of Scientific Instrument, 2014, 35(11): 2624-2632.
[3] Tang G, Luo G, Zhang W, et al. Underdetermined Blind Source Separation with Variational Mode Decomposition for Compound Roller Bearing Fault Signals[J]. Sensors, 2016, 16(6): 897.
[4] Wang X, Han W, Gu H, et al. Research on rolling element bearing fault diagnosis based on EEMD and correlated kurtosis[J]. Applied Mechanics and Materials, 2014, 680: 198-205.
[5] 丁康，黄志东，林慧斌. 一种谱峭度和Morlet小波的滚动轴承微弱故障诊断方法[J]. 振动工程学报，2014，27(01)：128-135.
Ding Kang, Huang Zhi-dong, Lin Hui-bin. A weak fault diagnosis method for rolling element bearings based on Morlet wavelet and spectral kurtosis[J]. Journal of Vibration Engineering, 2014, 27(01): 128-135.
[6] 林近山，陈前. 基于非平稳时间序列双标度指数特征的齿轮箱故障诊断[J]. 机械工程学报，2012，48(13)：108-114.
Lin Jin-shan, Chen Qian. Fault diagnosis of gearboxes based on the double-scaling-exponent characteristic of nonstationary time series[J]. Journal of Mechanical Engineering, 2012, 48(13): 108-114.
[7] 江星星，李舜酩，李世勋等. 泵车排量检测的极值增量DFA方法[J]. 振动.测试与诊断，2016，36(02)：227-232 +397.
Jiang Xing-xing, Li Shun-ming, Li Shi-xun, et al. A method for the discharge measurement of concrete pump based on DFA increment serious of extreme point[J]. Journal of Vibration, Measurement & Diagnosis, 2016, 36(02): 227-232 +397.
[8] 吴忠强，康晓华，于丹琦. 基于小波消噪和优化支持向量机的板形模式识别[J]. 中国机械工程，2018，29(01)：95-103.
Wu Zhong-qiang, Kang Xiao-hua, Yu Dan-qi. Flatness pattern recognition based on wavelet de-nosing and optimized SVM[J]. China Mechanical Engineering, 2018, 29(01): 95-103.
[9] Wang B, Zhang X, Fu Y, et al. Optimization of support vector machine and its application in intelligent fault diagnosis[J]. Journal of Vibration, Measurement and Diagnosis, 2017, 37(3): 547-552.
[10] Akbari M, Homaei H, Heidari M. An intelligent fault diagnosis approach for gears and bearings based on wavelet transform as a preprocessor and artificial neural networks[J]. International Journal of Mathematical Modelling & Computations, 2014, 4(4): 309-329.
[11] She M, Yang H, She M. Fault diagnosis method based on embedded acquisition equipment and SVM[J]. Advanced Materials Research, 2012, 468: 1681-1684.
[12] 王东风，孟丽，赵文杰. 基于自适应搜索中心的骨干粒子群算法[J]. 计算机学报，2016，39(12)：2652-2667.
Wang Dong-feng, Meng Li, Zhao Wen-jie. Improved bare bones particle swarm optimization with adaptive search center[J]. Chinese Journal of Computers, 2016, 39(12): 2652-2667.
[13] 黄泽霞，俞攸红，黄德才. 惯性权自适应调整的量子粒子群优化算法[J]. 上海交通大学学报，2012，46(2)：228-232.
Huang Ze-xia, Yu You-hong, Huang De-cai. Quantum-behaved particle swarm algorithm with self-adapting adjustment of inertia weight[J]. Journal of Shanghai Jiaotong University, 2012, 46(2): 228-232.
[14] Qian W, Li S, Wang J, et al. A novel supervised sparse feature extraction method and its application on rotating machine fault diagnosis[J]. Neurocomputing, 2018, 320: 129-140.
[15] Qian W, Li S, Wang J, et al. An intelligent fault diagnosis framework for raw vibration signals: adaptive overlapping convolutional neural network[J]. Measurement Science And Technology, 2018, 29(9): 095009.
[16] Xin Y, Li S, Wang J. A novel roller bearing fault diagnosis method based on the wavelet extreme learning machine[C]. 2017 Prognostics and System Health Management Conference, IEEE, 2017: 504-509
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
PDF(1054 KB)
