Accurate bearing fault diagnosis under varying operating conditions remains crucial for industrial reliability; however, the scarcity of labeled target data continues to significantly hinder the generalization of deep learning models. Existing unsupervised domain adaptation techniques often exhibit negative transfer by failing to distinguish between domain-invariant and domain-specific knowledge, leading to suboptimal adaptation. Therefore, this study proposes the Multisource Disentangled Expert Adaptation Network (MDEA-Net). MDEA-Net employs distinct shared expert networks to capture common fault characteristics across multiple source domains and private expert networks to model source-specific knowledge. Importantly, it explicitly disentangles these representations using an expert disentanglement loss, which raises both diversity and orthogonality. For efficient target adaptation, MDEA-Net utilizes frozen shared experts and introduces lightweight, target-specific private experts using low-rank adaptation. Domain alignment is accomplished by minimizing the maximum mean discrepancy specifically between shared feature spaces. Extensive experiments on benchmark datasets (Case Western Reserve University, Jiangnan University, Xi’an Jiaotong University and Changxing Sumyoung Technology Co., Ltd) demonstrate that MDEA-Net significantly outperforms state-of-the-art methods in cross-domain bearing fault diagnosis.
DingYMaLMaJ, et al. Intelligent fault diagnosis for rotating machinery using deep Q-network based health state classification: a deep reinforcement learning approach. Adv Eng Inform2019; 42: 100977.
2.
RandallRBAntoniJ.Rolling element bearing diagnostics—a tutorial. Mech Syst Signal Process2011; 25(2): 485–520.
3.
ZhangSZhangSWangB, et al. Deep learning algorithms for bearing fault diagnostics—a comprehensive review. IEEE Access2020; 8: 29857–29881.
4.
PandiyanMBabuTN.Systematic review on fault diagnosis on rolling-element bearing. J Vib Eng Technol2024; 12(7): 8249–8283.
5.
KankarPKSharmaSCHarshaSP.Fault diagnosis of ball bearings using continuous wavelet transform. Appl Soft Comput2011; 11(2): 2300–2312.
6.
LeiY.Fault diagnosis of rotating machinery based on empirical mode decomposition. Struct Health Monit2017; 26: 259–292.
7.
KablaAMokraniK.Bearing fault diagnosis using Hilbert-Huang transform (HHT) and support vector machine (SVM). Mech Ind2016; 17(3): 308.
8.
LeiYLiNGuoL, et al. Machinery health prognostics: a systematic review from data acquisition to RUL prediction. Mech Syst Signal Process2018; 104: 799–834.
9.
WenLLiXGaoL, et al. A new convolutional neural network-based data-driven fault diagnosis method. IEEE Trans Ind Electron2017; 65(7): 5990–5998.
10.
LiuHZhouJZhengY, et al. Fault diagnosis of rolling bearings with recurrent neural network-based autoencoders. ISA Trans2018; 77: 167–178.
11.
JiaFLeiYLinJ, et al. Deep neural networks: a promising tool for fault characteristic mining and intelligent diagnosis of rotating machinery with massive data. Mech Syst Signal Process2016; 72: 303–315.
12.
WenLGaoLLiX.A new deep transfer learning based on sparse auto-encoder for fault diagnosis. IEEE Trans Syst Man Cybern Syst2017; 49(1): 136–144.
13.
PanSJYangQ.A survey on transfer learning. IEEE Trans Knowl Data Eng2009; 22(10): 1345–1359.
14.
LongMCaoYWangJ, et al. Learning transferable features with deep adaptation networks. In: International conference on machine learning, Lille, France, 6–11 July 2015, pp. 97–105. London, UK: PMLR.
15.
SchwendemannSAmjadZSikoraA.Bearing fault diagnosis with intermediate domain based layered maximum mean discrepancy: a new transfer learning approach. Eng Appl Artif Intell2021; 105: 104415.
16.
GaninYUstinovaEAjakanH, et al. Domain-adversarial training of neural networks. J Mach Learn Res2016; 17: 2096–2030.
17.
ChenZHeGLiJ, et al. Domain adversarial transfer network for cross-domain fault diagnosis of rotary machinery. IEEE Trans Instrum Meas2020; 69(11): 8702–8712.
18.
JiangJZhaiC. Instance weighting for domain adaptation in NLP. Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics ACL, 2007, pp. 264–271.
19.
LinCZhaoSMengL, et al. Multi-source domain adaptation for visual sentiment classification. In: Proceedings of the AAAI conference on artificial intelligence, New York, NY, 7–12 February 2020, pp. 2661–2668. Palo Alto, CA: AAAI Press.
20.
ZhangWDengLZhangL, et al. A survey on negative transfer. IEEE-CAA J Autom Sinica2022; 10(2): 305–329.
21.
BousmalisKTrigeorgisGSilbermanN, et al. Domain separation networks. Adv Neural Inf Process Syst2016; 29: 343–351.
22.
YouKLongMCaoZ, et al. Universal domain adaptation. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, Long Beach, CA, 15–20 June 2019, pp. 2720–2729. Los Alamitos, CA: IEEE Computer Society.
23.
HuEJShenYWallisP, et al. Lora: low-rank adaptation of large language models. ICLR2022; 1(2): 3.
24.
DyerDStewartR.Detection of rolling element bearing damage by statistical vibration analysis. J Mech Des1979; 101(2): 229–235.
25.
TandonNChoudhuryA.A review of vibration and acoustic measurement methods for the detection of defects in rolling element bearings. Tribol Int1999; 32(8): 469–480.
26.
SaruhanHSaridemirSQicekA, et al. Vibration analysis of rolling element bearings defects. J Appl Res Technol2014; 12(3): 384–395.
27.
MoshrefzadehA.Condition monitoring and intelligent diagnosis of rolling element bearings under constant/variable load and speed conditions. Mech Syst Signal Process2021; 149: 107153.
28.
SmithWARandallRB.Rolling element bearing diagnostics using the Case Western Reserve University data: a benchmark study. Mech Syst Signal Process2015; 64: 100–131.
29.
JayaswalPVermaSNWadhwaniAK.Application of ANN, fuzzy logic and wavelet transform in machine fault diagnosis using vibration signal analysis. J Qual Maint Eng2010; 16(2): 190–213.
30.
QiuHLeeJLinJ, et al. Wavelet filter-based weak signature detection method and its application on rolling element bearing prognostics. J Sound Vib2006; 289(4–5): 1066–1090.
31.
FengZLiangMChuF.Recent advances in time–frequency analysis methods for machinery fault diagnosis: a review with application examples. Mech Syst Signal Process2013; 38(1): 165–205.
32.
HuangNEShenZLongSR, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis. Proc R Soc A1998; 454(1971): 903–995.
33.
JunshengCDejieYYuY.A fault diagnosis approach for roller bearings based on EMD method and AR model. Mech Syst Signal Process2006; 20(2): 350-362.
34.
TyagiCS.A comparative study of SVM classifiers and artificial neural networks application for rolling element bearing fault diagnosis using wavelet transform preprocessing. Neuron2008; 1: 309–317.
35.
SugumaranVSabareeshGRamachandranK.Fault diagnostics of roller bearing using kernel based neighborhood score multi-class support vector machine. Expert Syst Appl2008; 34(4): 3090–3098.
36.
JackLBNandiAK.Fault detection using support vector machines and artificial neural networks, augmented by genetic algorithms. Mech Syst Signal Process2002; 16(2–3): 373–390.
37.
SahriZBYusofRB.Support vector machine-based fault diagnosis of power transformer using k nearest-neighbor imputed DGA dataset. J Comput Commun2014; 2(9): 22–31.
38.
KankarPKSharmaSCHarshaSP.Fault diagnosis of ball bearings using machine learning methods. Expert Syst Appl2011; 38(3): 1876–1886.
39.
VishwendraMASalunkhePSPatilSV, et al. A novel method to classify rolling element bearing faults using K-nearest neighbor machine learning algorithm. ASCE ASME J Risk Uncertain Eng Syst Part B Mech Eng2022; 8(3): 031202.
40.
AherwarA.An investigation on gearbox fault detection using vibration analysis techniques: a review. Aust J Mech Eng2012; 10(2): 169–183.
41.
SamantaBAl-BalushiK.Artificial neural network based fault diagnostics of rolling element bearings using time-domain features. Mech Syst Signal Process2003; 17(2): 317–328.
42.
SakthivelNSugumaranVNairBB.Comparison of decision tree-fuzzy and rough set-fuzzy methods for fault categorization of mono-block centrifugal pump. Mech Syst Signal Process2010; 24(6): 1887–1906.
43.
ZhaoRYanRChenZ, et al. Deep learning and its applications to machine health monitoring. Mech Syst Signal Process2019; 115: 213–237.
44.
ZhangWLiCPengG, et al. A deep convolutional neural network with new training methods for bearing fault diagnosis under noisy environment and different working load. Mech Syst Signal Process2018; 100: 439–453.
45.
KayaYKuncanFErtunçHM.A new automatic bearing fault size diagnosis using time-frequency images of CWT and deep transfer learning methods. Turk J Electr Eng Comput Sci2022; 30(5): 1851–1867.
46.
JanssensOSlavkovikjVVervischB, et al. Convolutional neural network based fault detection for rotating machinery. J Sound Vib2016; 377: 331–345.
47.
VerstraeteDFerradaADroguettEL, et al. Deep learning enabled fault diagnosis using time-frequency image analysis of rolling element bearings. Shock Vib2017; 2017(1): 5067651.
48.
YuJZhangY.Challenges and opportunities of deep learning-based process fault detection and diagnosis: a review. Neural Comput Appl2023; 35(1): 211–252.
49.
Encalada-DávilaÁMoyónLTutivenC, et al. Early fault detection in the main bearing of wind turbines based on gated recurrent unit (GRU) neural networks and SCADA data. IEEE ASME Trans Mechatron2022; 27(6): 5583–5593.
50.
JiangGHeHYanJ, et al. Multiscale convolutional neural networks for fault diagnosis of wind turbine gearbox. IEEE Trans Ind Electron2018; 66(4): 3196–3207.
51.
ZhuJJiangQShenY, et al. Application of recurrent neural network to mechanical fault diagnosis: a review. J Mech Sci Technol2022; 36(2): 527–542.
52.
ZhaoWWangZLuC, et al. Fault diagnosis for centrifugal pumps using deep learning and softmax regression. In: 2016 12th world congress on intelligent control and automation (WCICA), Guilin, China, June 12–15, 2016, pp. 165–169. Piscataway, NJ: IEEE.
53.
ShiPGuoXHanD, et al. A sparse auto-encoder method based on compressed sensing and wavelet packet energy entropy for rolling bearing intelligent fault diagnosis. J Mech Sci Technol2020; 34(4): 1445–1458.
54.
MaZ-LLiX-JNianF-Q.An interpretable fault detection approach for industrial processes based on improved autoencoder. In: IEEE Transactions on Instrumentation and Measurement, 2025, vol. 74, pp. 1–13. IEEE.
55.
MaMSunCChenX.Deep coupling autoencoder for fault diagnosis with multimodal sensory data. IEEE Trans Industr Inform2018; 14(3): 1137–1145.
56.
ZhangYZhouTHuangX, et al. Fault diagnosis of rotating machinery based on recurrent neural networks. Measurement2021; 171: 108774.
57.
MaSChuFHanQ.Deep residual learning with demodulated time-frequency features for fault diagnosis of planetary gearbox under nonstationary running conditions. Mech Syst Signal Process2019; 127: 190–201.
58.
LiangPDengCWuJ, et al. Intelligent fault diagnosis of rotating machinery via wavelet transform, generative adversarial nets and convolutional neural network. Measurement2020; 159: 107768.
59.
ZhaoBZhangXZhanZ, et al. Deep multi-scale convolutional transfer learning network: a novel method for intelligent fault diagnosis of rolling bearings under variable working conditions and domains. Neurocomputing2020; 407: 24–38.
60.
TzengEHoffmanJSaenkoK, et al. Adversarial discriminative domain adaptation. In: Proceedings of the IEEE conference on computer vision and pattern recognition, Honolulu, HI, 21–26 July 2017, pp. 7167–7176.
61.
GrettonABorgwardtKMRaschMJ, et al. A kernel two-sample test. J Mach Learn Res2012; 13(1): 723–773.
62.
YangBLeiYJiaF, et al. A polynomial kernel induced distance metric to improve deep transfer learning for fault diagnosis of machines. IEEE Trans Ind Electron2019; 67(11): 9747–9757. Los Alamitos, CA: IEEE Computer Society.
63.
SunBFengJSaenkoK. Return of frustratingly easy domain adaptation. In: Proceedings of the AAAI conference on artificial intelligence, Phoenix, AZ, 12–17 February 2016. Palo Alto, CA: AAAI Press.
64.
LiXZhangWMaH, et al. Deep learning-based adversarial multi-classifier optimization for cross-domain machinery fault diagnostics. J Manuf Syst2020; 55: 334–347.
65.
GhifaryMKleijnWBZhangM, et al. Deep reconstruction-classification networks for unsupervised domain adaptation. In: European conference on computer vision, Amsterdam, the Netherlands, 8–16 October 2016, pp. 597–613. Switzerland: Springer, Cham.
66.
ChangQFangCZhouW, et al. A multi-order moment matching-based unsupervised domain adaptation with application to cross-working condition fault diagnosis of rolling bearings. Struct Health Monit2025; 24(3): 1438–1455.
67.
ChangQFangCFanL, et al. A dynamic weighted joint distribution domain adaptation network for cross-machine fault diagnosis of rolling bearings. Struct Health Monit2025; 24: 1–16.
68.
XuHWangSYuX, et al. SpecCSAM: a prior-knowledge embedding deep learning fault diagnosis method for unknown working conditions. Struct Health Monit2025; 24: 1–15.
69.
JacobsRAJordanMINowlanSJ, et al. Adaptive mixtures of local experts. Neural Comput1991; 3(1): 79–87.
70.
ShazeerNMirhoseiniAMaziarzK, et al. Outrageously large neural networks: the sparsely-gated mixture-of-experts layer. arXiv:1701.06538, 2017.
71.
MaJZhaoZYiX, et al. Modeling task relationships in multi-task learning with multi-gate mixture-of-experts. In: Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining, London, UK, 19–23 August 2018, pp. 1930–1939. New York, NY: ACM.
72.
HanBZhengCChanH, et al. Learning domain invariant representations in goal-conditioned block MDPS. Adv Neural Inf Process Syst2021; 34: 764–776.
73.
EigenDRanzatoMASutskeverI. Learning factored representations in a deep mixture of experts. In: International conference on learning representations, Scottsdale, AZ, 2–4 May 2013. London, UK: PMLR.
74.
TishbyNPereiraFCBialekW. The information bottleneck method. arXiv:physics/0004057, 2000.
75.
AdamKDBJ. A method for stochastic optimization. arXiv:1412.6980, 2014.
MagarRGhuleLLiJ, et al. FaultNet: a deep convolutional neural network for bearing fault classification. IEEE Access2021; 9: 25189–25199.
80.
LongMCaoZWangJ, et al. Conditional adversarial domain adaptation. Adv Neural Inf Process Syst2018; 31: 1640–1650.
81.
ChenXWangSLongM, et al. Transferability vs. discriminability: batch spectral penalization for adversarial domain adaptation. In: International conference on machine learning, Long Beach, CA, 9–15 June 2019, pp. 1081–1090. London, UK: PMLR.
82.
XuRLiGYangJ, et al. Larger norm more transferable: an adaptive feature norm approach for unsupervised domain adaptation. In: Proceedings of the IEEE/CVF international conference on computer vision, Seoul, Republic of Korea, 27 October–2 November 2019, pp. 1426–1435. Los Alamitos, CA: IEEE Computer Society.
83.
ShiYDengADengM, et al. A novel multiscale feature adversarial fusion network for unsupervised cross-domain fault diagnosis. Measurement2022; 200: 111616.
84.
ZhangYRenZZhouS, et al. Adversarial domain adaptation with classifier alignment for cross-domain intelligent fault diagnosis of multiple source domains. Meas Sci Technol2020; 32(3): 035102.
85.
LiDYangYSongY-Z, et al. Learning to generalize: Meta-learning for domain generalization. In: Proceedings of the AAAI conference on artificial intelligence, New Orleans, LA, 2–7 February 2018. Palo Alto, CA: AAAI Press.
86.
ExtrapolationVR.Out-of-distribution generalization via risk extrapolation. In: International conference on machine learning, Virtual Event, 18–24 July 2021, pp. 5815–5826. London, UK: PMLR.
87.
JiaNHuangWChengY, et al. A cross-domain intelligent fault diagnosis method based on multi-source domain feature adaptation and selection. Meas Sci Technol2024; 35(4): 046108.
88.
XuYChenLZhangF, et al. Multi-source alignment domain adaptation with similarity measurement for cross-domain bearing fault diagnosis. Meas Sci Technol2023; 34(5): 055006.
89.
ZhangYZhangSZhuY, et al. Cross-domain bearing fault diagnosis using dual-path convolutional neural networks and multi-parallel graph convolutional networks. ISA Trans2024; 152: 129–142.
90.
BiYFuRJiangC, et al. Single-source cross-domain bearing fault diagnosis via multipseudo-domain-augmented adversarial domain-invariant learning. IEEE Internet Things J2024; 11(19): 31521–31533.
91.
HouYMiJLuJ, et al. A deep transferable convolution neural network for unsupervised cross-domain diagnostics of rolling bearings in industrial internet of things. IEEE Internet Things J2025; 12(24): 20815–20826.
92.
WuSShiPXuX, et al. KMDSAN: a novel method for cross-domain and unsupervised bearing fault diagnosis. Knowl-Based Syst2025; 312: 113170.
93.
PaszkeAGrossSMassaF, et al. Pytorch: an imperative style, high-performance deep learning library. Adv Neural Inf Process Syst2019; 32: 8015–8026.
94.
MaatenLHintonG.Visualizing data using t-SNE. J Mach Learn Res2008; 9(2): 2579–2605.
