Effective and reliable health monitoring is critical for ensuring the operational safety, efficiency, and longevity of industrial machinery. Multi-layer perceptron (MLP)-based neural networks, referred to as “Universal Approximators,” often face challenges in handling complex, non-linear data due to high computational costs, limited continual learning, and reliance on fixed, non-learnable activation functions. To address these challenges, this article evaluates the potential of Kolmogorov-Arnold networks (KANs) to enhance efficiency and generalization in machinery anomaly detection using a triaxial vibration data across four different health states. Unlike MLPs, KAN employs spline-based, trainable activations on edges, enhancing flexibility and learning efficiency. This architecture improves classification accuracy, outperforming MLP-based neural networks, and conventional autoencoders by 5.2%, 10.64%, and 6.54% on X, Y, and Z-axis data, respectively. Beyond accuracy, KAN demonstrates faster training, superior generalization, improved continual learning, and resilience to catastrophic forgetting. This study establishes KAN as a promising framework for next-generation intelligent machine health monitoring systems for real-world engineering applications such as predictive maintenance, condition-based monitoring, and autonomous fault diagnosis in rotating equipment, spindles in machining operation, wind turbine gearboxes, and industrial fans used in steel and power plants.
CawleyP.Structural health monitoring: closing the gap between research and industrial deployment. Struct Health Monit2018; 17: 1225–1244.
2.
YuanF-GZargarSAChenQ, et al. Machine learning for structural health monitoring: challenges and opportunities. In: Sensors and smart structures technologies for civil, mechanical, and aerospace systems, California, United States, 2020, vol. 11379, p. 1137903.
3.
TianJMorilloCAzarianMH, et al. Motor bearing fault detection using spectral Kurtosis-based feature extraction coupled with K-nearest neighbor distance analysis. IEEE Trans Ind Electron2016; 63: 1793–1803.
4.
HanTJiangDZhaoQ, et al. Comparison of random forest, artificial neural networks and support vector machine for intelligent diagnosis of rotating machinery. Trans Inst Meas Control2017; 40: 2681–2693.
5.
LiuZGuoWHuJ, et al. A hybrid intelligent multi-fault detection method for rotating machinery based on RSGWPT, KPCA and Twin SVM. ISA Trans2017; 66: 249–261.
6.
ZhaoRYanRChenZ, et al. Deep learning and its applications to machine health monitoring. Mech Syst Signal Process2019; 115: 213–237.
7.
KelekoATKamsu-FoguemBNgounaRH, et al. Health condition monitoring of a complex hydraulic system using deep neural network and DeepSHAP explainable XAI. Adv Eng Softw2023; 175: 103339.
8.
BaiRNomanKFengK, et al. A two-phase-based deep neural network for simultaneous health monitoring and prediction of rolling bearings. Reliab Eng Syst Saf2023; 238: 109428.
9.
XuXCaoDZhouY, et al. Application of neural network algorithm in fault diagnosis of mechanical intelligence. Mech Syst Signal Process2020; 141: 106625.
10.
WangRChenZZhangS, et al. Dual-attention generative adversarial networks for fault diagnosis under the class-imbalanced conditions. IEEE Sens J2022; 22: 1474–1485.
11.
TamilselvanPWangYWangP. Deep belief network based state classification for structural health diagnosis. In: 2012 IEEE Aerospace Conference, 2012, pp. 1–11.
12.
TamilselvanPWangP.Failure diagnosis using deep belief learning based health state classification. Reliab Eng Syst Saf2013; 115: 124–135.
13.
MaMSunCChenX.Discriminative deep belief networks with Ant colony optimization for health status assessment of machine. IEEE Trans Instrum Meas2017; 66: 3115–3125.
14.
XingSLeiYWangS, et al. Distribution-invariant deep belief network for intelligent fault diagnosis of machines under new working conditions. IEEE Trans Ind Electron2021; 68: 2617–2625.
15.
ChenFLiuLTangB, et al. A novel fusion approach of deep convolution neural network with auto-encoder and its application in planetary gearbox fault diagnosis. Proc Inst Mech Eng Part O J Risk Reliab2020; 235: 3–16.
16.
ShaoHJiangHZhaoH, et al. A novel deep autoencoder feature learning method for rotating machinery fault diagnosis. Mech Syst Signal Process2017; 95: 187–204.
17.
LiYZengYQingY, et al. Learning local discriminative representations via extreme learning machine for machine fault diagnosis. Neurocomputing2020; 409: 275–285.
18.
HuangYChenC-HHuangC-J.Motor fault detection and feature extraction using RNN-based variational autoencoder. IEEE Access2019; 7: 139086–139096.
19.
JangG-BChoS-B.Feature space transformation for fault diagnosis of rotating machinery under different working conditions. Sensors2021; 21: 1417.
20.
LiYZouWJiangL.Fault diagnosis of rotating machinery based on combination of Wasserstein generative adversarial networks and long short term memory fully convolutional network. Measurement2022; 191: 110826.
21.
ChenXZhangBGaoD.Bearing fault diagnosis base on multi-scale CNN and LSTM model. J Intell Manuf2021; 32: 971–987.
22.
DurbhakaGKSelvarajBMittalM, et al. Swarm-LSTM: condition monitoring of gearbox fault diagnosis based on hybrid LSTM deep neural network optimized by swarm intelligence algorithms. Comput Mater Contin2021; 66: 2041–2059.
23.
ChengYQvJFengK, et al. A Bayesian adversarial probsparse transformer model for long-term remaining useful life prediction. Reliab Eng Syst Saf2024; 248: 110188.
24.
WangHLiSSongL, et al. A novel convolutional neural network based fault recognition method via image fusion of multi-vibration-signals. Comput Ind2019; 105: 182–190.
25.
QiuGGuYCaiQ.A deep convolutional neural networks model for intelligent fault diagnosis of a gearbox under different operational conditions. Measurement2019; 145: 94–107.
26.
AyankosoSDuttaAHeY, et al. Performance of vibration and current signals in the fault diagnosis of induction motors using deep learning and machine learning techniques. Struct Health Monit. Epub ahead of print 8 November 2024. DOI: 10.1177/14759217241289874.
27.
HanYTangBDengL.An enhanced convolutional neural network with enlarged receptive fields for fault diagnosis of planetary gearboxes. Comput Ind2019; 107: 50–58.
28.
HanTLiY-FQianM.A hybrid generalization network for intelligent fault diagnosis of rotating machinery under unseen working conditions. IEEE Trans Instrum Meas2021; 70: 1–11.
29.
ZhangXWangCZhouW, et al. Trustworthy diagnostics with out-of-distribution detection: a novel max-consistency and min-similarity guided deep ensembles for uncertainty estimation. IEEE Internet Things J2024; 11: 23055–23067.
30.
FengKXiaoHZhangJ, et al. A digital twin methodology for vibration-based monitoring and prediction of gear wear. Wear2025; 571: 205806.
31.
FengKJiJCZhangY, et al. Digital twin-driven intelligent assessment of gear surface degradation. Mech Syst Signal Process2023; 186: 109896.
32.
FengKXuYWangY, et al. Digital twin enabled domain adversarial graph networks for bearing fault diagnosis. IEEE Trans Ind Cyber Phys Syst2023; 1: 113–122.
33.
LiSJiJFengK, et al. Composite neuro-fuzzy system-guided cross-modal zero-sample diagnostic framework using multisource heterogeneous noncontact sensing data. IEEE Trans Fuzzy Syst2025; 33: 302–313.
34.
LiuZWangYVaidyaS, et al. KAN: Kolmogorov-Arnold networks. arXiv Prepr arXiv240419756. Epub ahead of print 2024. DOI: 10.48550/arXiv.2404.19756.
KoenigBCKimSDengS.KAN-ODEs: Kolmogorov–Arnold network ordinary differential equations for learning dynamical systems and hidden physics. Comput Methods Appl Mech Eng2024; 432: 117397.
42.
Al-qanessMAANiS. TCNN-KAN: optimized CNN by Kolmogorov-Arnold network and pruning techniques for sEMG Gesture recognition. IEEE J Biomed Heal Informatics2024; 29(1): 188–197.
43.
GuiSWangJTangC, et al. High precision frequency source anomaly monitor using Kolmogorov-Arnold networks. In: 2024 22nd International conference on optical communications and networks (ICOCN), Harbin, China, 2024, pp. 1–2.
44.
PangLLiJGanJ, et al. Dissolved gas analysis of transformers based on Kolmogorov-Arnold network. In: 2024 6th International conference on energy systems and electrical power (ICESEP), Wuhan, China, 2024, pp. 1568–1571.
45.
KingmaDPBaJL. Adam: a method for stochastic optimization. In: 3rd International conference on learning representations, ICLR 2015 - Conference Track Proceedings. International Conference on Learning Representations, ICLR, 2014.
46.
SmithLN.A disciplined approach to neural network hyper-parameters: Part 1 - learning rate, batch size, momentum, and weight decay. arXiv Prepr arXiv.1803. 09820. Epub ahead of print 2018. DOI: 10.48550/arXiv.1803.09820.
47.
NayakSBhatMSubba ReddyNV, et al. Study of distance metrics on K-nearest neighbor algorithm for star categorization. J Phys Conf Ser2022; 2161: 12004.
48.
MypatiOMukherjeeAMishraD, et al. A critical review on applications of artificial intelligence in manufacturing. Artif Intell Rev2023; 56: 661–768.
49.
CybenkoG.Approximation by superpositions of a sigmoidal function. Math Control Signals Syst1989; 2: 303–314.
50.
LeshnoMLinVYPinkusA, et al. Multilayer feedforward networks with a nonpolynomial activation function can approximate any function. Neural Netw1993; 6: 861–867.
51.
GuliyevNJIsmailovVE.On the approximation by single hidden layer feedforward neural networks with fixed weights. Neural Netw2018; 98: 296–304.
52.
SunYMaoHGuoQ, et al. Learning a good representation with unsymmetrical auto-encoder. Neural Comput Appl2016; 27: 1361–1367.
53.
ZhangCBengioSHardtM, et al. Understanding deep learning requires rethinking generalization. arXiv.1611. 03530. Epub ahead of print 2016. DOI: 10.48550/arXiv. 1611.03530.
54.
McCloskeyMCohenNJ.Catastrophic interference in connectionist networks: the sequential learning problem. Psychol Learn Motivation1989; 24: 109–165.
55.
MermillodMBugaiskaABoninP.The stability-plasticity dilemma: investigating the continuum from catastrophic forgetting to age-limited learning effects. Front Psychol2013; 4: 504.
