摘要由于滚动轴承实际工作环境恶劣,含标签故障样本数据严重缺乏,不足以建立准确的预测模型。支持矩阵机(Support Matrix Machine, SMM)作为一种新的模式识别方法,可以获得良好的分类效果,但其仍对小样本分析具有局限性。基于此,论文提出一种迁移最小二乘支持矩阵机(Transfer Least Square Support Matrix Machine,TLSSMM)分类方法。在TLSSMM分类过程中,利用源域样本训练得到近似目标域的预测模型,并通过目标域少量含标签样本微调源域的训练模型以更新得到新模型。同时,采用最小二乘损失来约束目标函数,使其由不等式转换为等式,只需求解一组线性方程即可获得结果,大大提升分类效率。选择两种不同的滚动轴承故障数据对所提方法进行验证,实验结果表明,TLSSMM方法具有优异的分类性能。
关键词:迁移最小二乘支持矩阵机;最小二乘损失;滚动轴承;故障诊断
Abstract:Due to the harsh working environment of rolling bearings, there is a serious lack of data on rolling bearings with labels, which is not enough to establish an accurate prediction model. Support Matrix Machine (SMM), as a new pattern recognition method, can obtain good classification results, but it still has limitations for small sample classification. Based on this, this paper proposes a transfer least squares support matrix machine classification (TLSSMM) method. In the classification process of TLSSMM, the source domain samples are used to train and obtain a prediction model that approximates the target domain, and a small number of labeled samples in the target domain are used to fine-tune the training model of the source domain to update the new model. At the same time, the least square loss is used to constrain the objective function, so that it is converted from an inequality to an equation, and the result can be obtained by solving a set of linear equations, which greatly improves the efficiency of classification. Two different rolling bearing fault data are selected to verify the proposed method. The experimental results show that the TLSSMM method has excellent classification performance.
Keywords: Transfer Least Squares Support Matrix Machine; Least Square Loss; Rolling Bearing; Fault Diagnosis
伍毅,盛丽,潘海洋,郑近德. 基于迁移最小二乘支持矩阵机的滚动轴承故障诊断方法[J]. 振动与冲击, 2022, 41(21): 53-59.
WU Yi, SHENG Li, PAN Haiyang, ZHENG Jinde. Fault diagnosis method of rolling bearing based on transfer least squares support matrix machine. JOURNAL OF VIBRATION AND SHOCK, 2022, 41(21): 53-59.
[1] 刘畅, 伍星, 刘韬,等. 基于近似等距投影和支持向量机的滚动轴承故障诊断[J]. 振动与冲击, 2018, 37(005):234-239.
Liu Chang, Wu Xing, Liu Tao, et al. Rolling bearing fault diagnosis based on approximate isometric projection and support vector machine [J]. Vibration and shock, 2018, 37 (005): 234-239
[2] Li Wendi, Ng Wing W Y, Wang Ting, et al. HELP: An LSTM-based approach to hyperparameter exploration in neural network learning[J]. Neurocomputing, 2021, 442: 161-172.
[3] Yu Anlan, Jing Shusen, Tan Xiaosi, et al. Efficient successive over relaxation detectors for massive MIMO[J]. IEEE Transactions on Circuits and Systems I: Regular Papers, 2020, 67(6): 2128-2139.
[4] Khemchandani R, Chandra S. Twin support vector machines for pattern classification[J]. IEEE Transactions on pattern analysis and machine intelligence, 2007, 29(5): 905-910.
[5] Luo Chuan, Huang Chi, Cao Jinde, et al. Short-term traffic flow prediction based on least square support vector machine with hybrid optimization algorithm[J]. Neural processing letters, 2019, 50(3): 2305-2322.
[6] Khalifa H A E W, Alharbi M, Kumar P. A new method for solving quadratic fractional programming problem in neutrosophic environment[J]. Open Engineering, 2021, 11(1): 880-886.
[7] Luo Luo, Xie Yubo, Zhang Zhihua, et al. Support matrix machines[C]//International conference on machine learning. PMLR, 2015: 938-947
[8] 巩朋成, 王兆彬, 谭海明,等. 杂波背景下基于交替方向乘子法的低截获频控阵MIMO雷达收发联合优化方法[J]. 电子与信息学报, 2021, 43(5):1267-1274.
Gong Pengcheng, Wang Zhaobin, Tan Haiming, et al. Joint transceiver optimization method of low intercept frequency controlled array MIMO Radar Based on alternating direction multiplier method in clutter background [J]. Journal of electronics and information, 2021, 43 (5): 1267-1274
[9] Zheng Qingqing, Zhu Fengyuan, Qin Jing, et al. Sparse support matrix machine[J]. Pattern Recognition, 2018, 76: 715-726.
[10] Pan Haiyang, Yang Yu, Zheng Jinde, et al. Symplectic interactive support matrix machine and its application in roller bearing condition monitoring[J]. Neurocomputing, 2020, 398: 1-10.
[11] Pan Haiyang, Zheng Jinde. An intelligent fault diagnosis method for roller bearing using symplectic hyperdisk matrix machine[J]. Applied Soft Computing, 2021,105:107284.
[12] Guo Yiqing, Jia Xiuping, Paull David. A domain-transfer support vector machine for multi-temporal remote sensing imagery classification[C]//2017 IEEE International Geoscience and Remote Sensing Symposium (IGARSS). IEEE, 2017: 2215-2218.
[13] Chen Yuqing, Huang Yunsong, Huang Lianjie. Suppressing migration image artifacts using a support vector machine method[J]. Geophysics, 2020, 85(5): S255-S268.
[14] Xu Dan, et al. Application of friedman test and multiple analytical method in preparation of liquid essence of Antarctic krill[J]. China Condiment, 2019, 44(1): 89-92.
[15] Richhariya B, Tanveer M. A robust fuzzy least squares twin support vector machine for class imbalance learning[J]. Applied Soft Computing, 2018, 71: 418-432.
[16] 程正阳, 王荣吉, 潘海洋. 辛几何模态分解方法及其分解能力研究[J]. 振动与冲击, 2020,39(13):27-35.
Cheng Zhengyang, Wang Rongji, pan Haiyang. Study on symplectic geometric mode decomposition method and its decomposition ability [J]. Vibration shock, 2020,39(13):27-35.