Abstract: Predicting remaining useful life (RUL) of bearings under scarce labeled data is significant for intelligent manufacturing. Current approaches typically encounter the challenge that different degradation stages have similar behaviors in multisensor scenarios. Given that cross-sensor similarity improves the discrimination of degradation features, we propose a multisensor contrast method for RUL prediction under scarce RUL-labeled data, in which we use cross-sensor similarity to mine multisensor similar representations that indicate machine health condition from rich unlabeled sensor data in a co-occurrence space. Specifically, we use ResNet18 to span the features of different sensors into the co-occurrence space. We then obtain multisensor similar representations of abundant unlabeled data through alternate contrast based on cross-sensor similarity in the co-occurrence space. The multisensor similar representations indicate the machine degradation stage. Finally, we focus on finetuning these similar representations to achieve RUL prediction with limited labeled sensor data. The proposed method is evaluated on a publicly available bearing dataset, and the results show that the mean absolute percentage error is reduced by at least 0.058, and the score is improved by at least 0.122 compared with those of state-of-the-art methods.

标签数据稀缺下基于多传感器对比神经网络的滚动轴承剩余使用寿命预测

[1]Akrim A, Gogu C, Vingerhoeds R, et al., 2023. Self-supervised learning for data scarcity in a fatigue damage prognostic problem. Eng Appl Artif Intell, 120: 105837.

[2]Behera S, Misra R, 2023. A multi-model data-fusion based deep transfer learning for improved remaining useful life estimation for IIOT based systems. Eng Appl Artif Intell, 119: 105712.

[3]Deng YF, Du SC, Wang D, et al., 2023. A calibration-based hybrid transfer learning framework for RUL prediction of rolling bearing across different machines. IEEE Trans Instrum Meas, 72:1-15.

[4]Ding YF, Zhuang JC, Ding P, et al., 2022. Self-supervised pretraining via contrast learning for intelligent incipient fault detection of bearings. Reliab Eng Syst Saf, 218: 108126.

[5]Ding YF, Jia MP, Cao YD, et al., 2023. Domain generalization via adversarial out-domain augmentation for remaining useful life prediction of bearings under unseen conditions. Knowl-Based Syst, 261: 110199.

[6]Dong SJ, Xiao JF, Hu XL, et al., 2023. Deep transfer learning based on Bi-LSTM and attention for remaining useful life prediction of rolling bearing. Reliab Eng Syst Saf, 230: 108914.

[7]He KM, Fan HQ, Wu YX, et al., 2020. Momentum contrast for unsupervised visual representation learning. Proc IEEE/CVF Conf on Computer Vision and Pattern Recognition, p.9729-9738.

[8]Kong ZQ, Jin XH, Xu ZG, et al., 2023. A contrastive learning framework enhanced by unlabeled samples for remaining useful life prediction. Reliab Eng Syst Saf, 234: 109163.

[9]Korbar B, Tran D, Torresani L, 2018. Cooperative learning of audio and video models from self-supervised synchronization. Proc 32^nd Int Conf on Neural Information Processing Systems, p.7774-7785.

[10]Krokotsch T, Knaak M, Gühmann C, et al., 2022. Improving semi-supervised learning for remaining useful lifetime estimation through self-supervision. Int J Progn Health Manage, 13(1):853.

[11]Li Q, Yan CF, Chen GY, et al., 2022. Remaining useful life prediction of rolling bearings based on risk assessment and degradation state coefficient. ISA Trans, 129:413-428.

[12]Li Y, Wang HJ, Li JW, et al., 2022. A 2-D long short-term memory fusion networks for bearing remaining useful life prediction. IEEE Sens J, 22(22):21806-21815.

[13]Melendez I, Doelling R, Bringmann O, 2019. Self-supervised multi-stage estimation of remaining useful life for electric drive units. Proc IEEE Int Conf on Big Data, p.4402-4411.

[14]Morales-Espejel GE, Gabelli A, 2020. A model for rolling bearing life with surface and subsurface survival: surface thermal effects. Wear, 460-461: 203446.

[15]Nectoux P, Gouriveau R, Medjaher K, et al., 2012. PRONOSTIA: an experimental platform for bearings accelerated degradation tests. Proc IEEE Int Conf on Prognostics and Health Management, p.1-8.

[16]Saeed A, Salim FD, Ozcelebi T, et al., 2021. Federated selfsupervised learning of multisensor representations for embedded intelligence. IEEE Int Things J, 8(2):1030-1040.

[17]Souza JS, Bezerril MC, Silva MA, et al., 2021. Motor speed estimation and failure detection of a small UAV using density of maxima. Front Inform Technol Electron Eng, 22(7):1002-1009.

[18]Tao F, Qi QL, Liu A, et al., 2018. Data-driven smart manufacturing. J Manuf Syst, 48:157-169.

[19]Tian YL, Krishnan D, Isola P, 2020. Contrastive multiview coding. Proc 16^th European Conf on Computer Vision, p.776-794.

[20]Wang B, Lei YG, Li NP, et al., 2020. Multiscale convolutional attention network for predicting remaining useful life of machinery. IEEE Trans Ind Electron, 68(8):7496-7504.

[21]Wang WJ, Wang Y, Wang J, et al., 2022. Ensemble enhanced active learning mixture discriminant analysis model and its application for semi-supervised fault classification. Front Inform Technol Electron Eng, 23(12):1814-1827.

[22]Wang X, Wang TY, Ming AB, et al., 2021. Cross-operating condition degradation knowledge learning for remaining useful life estimation of bearings. IEEE Trans Instrum Meas, 70: 3520911.

[23]Wang YT, Cai F, Pan ZQ, et al., 2023. Self-supervised graph learning with target-adaptive masking for session-based recommendation. Front Inform Technol Electron Eng, 24(1):73-87.

[24]Wen YX, Wu JG, Zhou Q, et al., 2019. Multiple-change-point modeling and exact Bayesian inference of degradation signal for prognostic improvement. IEEE Trans Autom Sci Eng, 16(2):613-628.

[25]Yang L, Liao YH, Duan RK, et al., 2023. A bidirectional recursive gated dual attention unit based RUL prediction approach. Eng Appl Artif Intell, 120: 105885.

[26]Zhang BM, Mao YF, Chen X, et al., 2021. Self-supervised learning advance fault diagnosis of rotating machinery. Proc 2^nd Int Conf on Neural Computing for Advanced Applications, p.319-332.

[27]Zhang WW, Chen DJ, Kong Y, 2021. Self-supervised joint learning fault diagnosis method based on three-channel vibration images. Sensors, 21(14):4774.

[28]Zhu SH, Pu J, 2021. A self-supervised method for treatment recommendation in sepsis. Front Inform Technol Electron Eng, 22(7):926-939.

[29]Zou WH, Lu ZQ, Hu ZY, et al., 2023. Remaining useful life estimation of bearing using deep multiscale window-based transformer. IEEE Trans Instrum Meas, 72: 3514211.

[30]Zuo T, Zhang K, Zheng Q, et al., 2023. A hybrid attentionbased multi-wavelet coefficient fusion method in RUL prognosis of rolling bearings. Reliab Eng Syst Saf, 237: 109337.