Deep learning has exhibited remarkable accuracy in fault diagnosis. However, its opaque nature complicates the interpretation of decision-making processes, thereby constraining its dependability and practicality in industrial settings. This paper introduces an interpretable graph neural network model named Multi-Wavelet Kernel Graph Sample and Aggregate Network (MWK-GraphSage) to address three limitations in current approaches: limited modeling capacity for non-Euclidean data, insufficient disentanglement of multi-fault characteristics, and one-dimensional interpretability assessment. The proposed model features a multi-wavelet kernel convolutional layer (MWKC) that integrates four wavelet basis functions (Sine, Cauchy, Bump, and Laplace) to extract multi-scale time–frequency attributes, thereby enhancing the capability to disentangle features. An Adaptive Wavelet Weight (AWW) allocation mechanism is devised to dynamically enhance feature fusion performance. Furthermore, various visualization techniques are employed to facilitate multi-dimensional interpretability analysis. Experimental results on the Southeast University gearbox dataset demonstrate that MWK-GraphSage achieves a fault diagnosis accuracy of 99%, surpassing comparative models such as ResGAT and WaveletCNN.
AbuhamdiaTSaiedTJohnB, et al. (2018) Laplace wavelet transform theory and applications. Journal of vibration and control24(9): 1600–1620.
2.
AgarwalCJohnsonNPawelczykMet al. (2022) Rethinking Stability for Attribution-based Explanations. arXiv:2203.06877 [cs.LG]. Epub ahead of print 14 March 2022. DOI: 10.48550/arXiv.2203.06877.
3.
AlekseevABobeA (2019) GaborNet: Gabor filters with learnable parameters in deep convolutional neural network. In: 2019 International Conference on Engineering and Telecommunication (EnT). IEEE.
4.
BauDZhouBKhoslaA, et al. (2017) Network dissection: quantifying interpretability of deep visual representations. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,Honolulu, Hawaii, USA, 216541–266549July 2017.
5.
BergoldPLasserC (2020) Fourier series windowed by a bump function. Journal of Fourier Analysis and Applications26: 1–28.
6.
BounemeurAChemachemaM (2021) Adaptive fuzzy fault-tolerant control using Nussbaum-type function with state-dependent actuator failures. Neural Computing & Applications33(1): 191–208.
7.
BounemeurAChemachemaM (2023) General fuzzy adaptive fault-tolerant control based on Nussbaum-type function with additive and multiplicative sensor and state-dependent actuator faults. Fuzzy Sets and Systems468: 108616.
8.
BounemeurAChemachemaM (2025) Robust fuzzy adaptive fault-tolerant control for a class of second-order nonlinear systems. International Journal of Adaptive Control and Signal Processing39(1): 15–30.
9.
BounemeurAChemachemaMEssounbouliN (2018) Indirect adaptive fuzzy fault-tolerant tracking control for MIMO nonlinear systems with actuator and sensor failures. ISA Transactions79: 45–61.
10.
ChangLBrancoP (2024) Embedding residuals in graph-based solutions: the E-ResSAGE and E-ResGAT algorithms. A case study in intrusion detection. Applied Intelligence54(8): 6025–6040.
11.
ChenQChenKKDongXJ, et al. (2024) An interpretable convolutional neural network for mechanical equipment fault diagnosis. Journal of Mechanical Engineering60(12): 65–76.
12.
ChengKYWangNShiWX, et al. (2020) Research progress on interpretability of deep learning. Journal of Computer Research and Development57(6): 1208–1217.
13.
GongJYHuanLXZhengXW (2022) Interpretability analysis methods for deep learning in image interpretation. Acta Geodetica et Cartographica Sinica51(6): 873–884.
14.
GrezmakJWangPGaoRX, et al. (2019) Explainable convolutional neural network for gearbox fault diagnosis. Procedia CIRP80: 476–481.
15.
HuangQYingZDongY, et al. (2022) GraphLIME: local interpretable model explanations for graph neural networks. IEEE Transactions on Knowledge and Data Engineering35(7): 6968–6972.
16.
JiangGJiaoJLinJ, et al. (2023a) An interpretable convolutional neural network with multi-wavelet kernel fusion for intelligent fault diagnosis. Journal of Manufacturing Systems70: 18–30.
17.
KeshunYGuangqiQYingkuiG (2024) Optimizing prior distribution parameters for probabilistic prediction of remaining useful life using deep learning. Reliability Engineering & System Safety242: 109793.
18.
KeshunYPuzhouWPengH, et al. (2025b) A sound-vibration physical-information fusion constraint-guided deep learning method for rolling bearing fault diagnosis. Reliability Engineering & System Safety253: 110556.
19.
KeshunYYingkuiGYanghuiL, et al. (2025a) A novel physical constraint-guided quadratic neural networks for interpretable bearing fault diagnosis under zero-fault sample.Nondestructive Testing and Evaluation24: 1–31.
20.
KeshunYYanghuiLQiuG, et al. (2025c) DTMPI-DIVR: a digital twins for multi-margin physical information via dynamic interaction of virtual and real sound-vibration signals for bearing fault diagnosis without real fault samples. Expert Systems with Applications292: 128592.
21.
LiTWangBYangY, et al. (2021) WaveletKernelNet: an interpretable deep neural network for industrial intelligent diagnosis. IEEE Transactions on Systems, Man, and Cybernetics: Systems52(4): 2302–2312.
22.
LiQKMaRXLiaoMF, et al. (2023a) A novel fault evaluation method based on nonlinear vibration features and Euclidean distance measurement for grid-like structures. Structural Health Monitoring22(6): 4131–4148.
23.
LiYLuYZhangR, et al. (2023b) Deep learning for energy efficient beamforming in MU-MISO networks: a GAT-based approach. IEEE Wireless Communications Letters12(7): 1264–1268.
24.
LiTSunCFinkO, et al. (2024) Filter-informed spectral graph wavelet networks for multiscale feature extraction and intelligent fault diagnosis. IEEE Transactions on Cybernetics54(1): 506–518.
25.
LinJJiaoJY (2023) Research advances and challenges in interpretable mechanical intelligent diagnosis. Journal of Mechanical Engineering59(20): 215–224.
26.
LiuXLiuXMZhangRL, et al. (2022) Securely computing the Manhattan distance under the malicious model and its applications. Applied Sciences12(22): 458–463.
27.
LotfiniaHNajafiAShidfarA (2021) The bi-Helmholtz equation with Cauchy conditions: ill-posedness and regularization methods. Inverse Problems in Science and Engineering29(1): 17–39.
28.
LundbergSMLeeSI (2017) A unified approach to interpreting model predictions. In: Proceedings of the 31st International Conference on Neural Information Processing Systems. NeurIPS, pp. 4768–4777.
29.
MoTJLiuFDengXX (2025) A group decision making model with non-reciprocal fuzzy preference relations based on Pearson correlation coefficient. Expert Systems with Applications276: 127095.
30.
ShaoSYanRLuY, et al. (2018) Highly accurate machine fault diagnosis using deep transfer learning. IEEE Transactions on Industrial Informatics15(4): 2446–2455.
31.
ShenYShaoHWangX, et al. (2024) Few-shot class-incremental learning for system-level fault diagnosis of wind turbine. IEEE/ASME Transactions on Mechatronics (1941–014X): 1–10. DOI: 10.1109/TMECH.2024.3490733.
32.
SpinelliIScardapaneSUnciniA (2022) A meta-learning approach for training explainable graph neural networks. IEEE Transactions on Neural Networks and Learning Systems35(4): 4647–4655.
33.
VijayaraghavanSMohapatraP (2023) Stability of explainable recommendation. In: Proceedings of the 17th ACM Conference on Recommender Systems, Singapore, 18947–22954September 2023.
34.
WangJHCaoWBZhouDY, et al. (2025) Fault diagnosis of rolling bearings under variable operating conditions based on MSF and I-InceptionNet. Journal of Huazhong University of Science and Technology (Nature Science Edition)53(5): 24–30.
35.
WuKChenYChenH, et al. (2022) Self-attention memory-augmented wavelet-CNN for anomaly detection. IEEE Transactions on Circuits and Systems for Video Technology33(3): 1374–1385.
36.
ZhaoZLiTAnB, et al. (2022) Model-driven deep unrolling: towards interpretable deep learning against noise attacks for intelligent fault diagnosis. ISA Transactions129: 644–662.
37.
ZhengXShiraniFWangTet al. (2023) Towards Robust Fidelity for Evaluating Explainability of Graph Neural Networks. arXiv:2310.01820 [cs.LG]. Epub ahead of print 3 October 2023; revised 17March2024 (v2). DOI: 10.48550/arXiv.2310.01820.
