Predicting the remaining useful life (RUL) of milling tools is crucial for maximizing tool utilization and reducing machining costs. However, accurately predicting RUL faces three main issues: large data variations across variable conditions, potential conflicts with physical laws, and estimating prediction uncertainty. To address these limitations, this paper proposes multi-dimensionally screened features and a physically modified attention network to predict the milling tool RUL under variable conditions. A multi-dimensionally screened features method is designed to obtain minimal variation across variable conditions by evaluating correlation and similarity. The screened features are integrated with physical constraints to construct a unified health indicator. Finally, the physics-based model and Wiener process jointly guide the attention network training, narrowing the parameter search space and enhancing RUL prediction accuracy. Compared to data-driven and hybrid-driven models, the proposed model achieves optimal prediction performance and quantifies prediction uncertainty under variable conditions, with the lowest MSE reaching 0.68 and MAPE within 10%, showing an average improvement of 35.08% over hybrid-driven methods.
SerinGSenerBOzbayogluAM, et al. Review of tool condition monitoring in machining and opportunities for deep learning. Int J Adv Manuf Technol2020; 109: 953–974.
2.
ShenYYangFHabibullahMS, et al. Predicting tool wear size across multi-cutting conditions using advanced machine learning techniques. J Intell Manuf2021; 32: 1753–1766.
3.
AnQTaoZXuX, et al. A data-driven model for milling tool remaining useful life prediction with convolutional and stacked LSTM network. Measurement2020; 154: 107461.
4.
HuYMiaoXSiY, et al. Prognostics and health management: a review from the perspectives of design, development and decision. Reliab Eng Syst Saf2022; 217: 108063.
5.
AstakhovVP.The assessment of cutting tool wear. Int J Mach Tools Manuf2004; 44: 637–647.
6.
YeHWanYRenX, et al. Validity analysis of generalized taylor regression formula by tool wear test for difficult-to-machine materials. Key Eng Mater2013; 589–590: 342–350.
7.
ZhuZZhuWHuangJ, et al. An intelligent monitoring system for robotic milling process based on transfer learning and digital twin. J Manuf Syst2025; 78: 433–443.
8.
AizpuruaJIStewartBGMcArthurSDJ, et al. Probabilistic forecasting informed failure prognostics framework for improved RUL prediction under uncertainty: a transformer case study. Reliab Eng Syst Saf2022; 226: 108676.
9.
LinKWangYChenY.A Kernel-density based semi-parametric stochastic degradation model with dependent increments. Mech Syst Signal Process2021; 161: 107978.
10.
ZhangZSiXHuC, et al. Degradation data analysis and remaining useful life estimation: a review on wiener-process-based methods. Eur J Oper Res2018; 271: 775–796.
11.
WeiwenPLiY-FMiJ, et al. Reliability of complex systems under dynamic conditions: a Bayesian multivariate degradation perspective. Reliab Eng Syst Saf2016; 153: 75–87.
12.
PanZBalakrishnanNSunQ, et al. Bivariate degradation analysis of products based on Wiener processes and copulas. J Stat Comput Simul2012; 83: 1–14.
13.
PengWLiY-FYangY-J, et al. Bivariate analysis of incomplete degradation observations based on inverse Gaussian processes and copulas. IEEE Trans Reliab2016; 65: 1–16.
14.
LiuKHuangS.Integration of data fusion methodology and degradation modeling process to improve prognostics. IEEE Trans Autom Sci Eng2016; 13: 344–354.
15.
LiuKChehadeASongC.Optimize the signal quality of the composite health index via data fusion for degradation modeling and prognostic analysis. IEEE Trans Autom Sci Eng2017; 14: 1504–1514.
16.
YanHLiuKZhangX, et al. Multiple sensor data fusion for degradation modeling and prognostics under multiple operational conditions. IEEE Trans Reliab2016; 65: 1416–1426.
17.
ZhaoRYanRChenZ.Deep learning and its applications to machine health monitoring. Mech Syst Signal Process2019; 115: 213–237.
18.
LiGXuSJiangR, et al. physics-informed inhomogeneous wear identification of end mills by online monitoring data. J Manuf Process2024; 132: 759–771.
19.
ZhangXLuXLiW, et al. Prediction of the remaining useful life of cutting tool using the hurst exponent and CNN-LSTM. Int J Adv Manuf Technol2021; 112: 2277–2299.
20.
García NietoPJGarcía-GonzaloESánchez LasherasF, et al. Hybrid PSO–SVM-based method for forecasting of the remaining useful life for aircraft engines and evaluation of its reliability. Reliab Eng Syst Saf2015; 138: 219–231.
21.
LiGYangLLeeC-G, et al. A Bayesian deep learning RUL framework integrating epistemic and aleatoric uncertainties. IEEE Trans Ind Electron2021; 68: 8829–8841.
22.
FalconAD’AgostinoGLanzO, et al. Neural turing machines for the remaining useful life estimation problem. Comput Ind2022; 143: 103762.
23.
ZhouJ-TZhaoXGaoJ.Tool remaining useful life prediction method based on LSTM under variable working conditions. Int J Adv Manuf Technol2019; 104: 4715–4726.
24.
LiuCZhuL.A two-stage approach for predicting the remaining useful life of tools using bidirectional long short-term memory. Measurement2020; 164: 108029.
25.
XuXWangJZhongB, et al. Deep learning-based tool wear prediction and its application for machining process using multi-scale feature fusion and channel attention mechanism. Measurement2021; 177: 109254.
26.
LiXLiuXYueC, et al. Data-model linkage prediction of tool remaining useful life based on deep feature fusion and wiener process. J Manuf Syst2024; 73: 19–38.
27.
LiGXuSSunL.Accurate tool wear and breakage monitoring method for milling process based on vision and laser sensor fusion. Meas Sci Technol2024; 35: 065001.
28.
LiuCNiJWanP.An accurate prediction method of multiple deterioration forms of tool based on multitask learning with low rank tensor constraint. J Manuf Syst2021; 58: 193–204.
29.
LinY-HChangLGuanL-X.Enhanced stochastic recurrent hybrid model for RUL predictions via semi-supervised learning. Reliab Eng Syst Saf2024; 248: 110167.
30.
ZhangZChenXZioE, et al. Multi-task learning boosted predictions of the remaining useful life of aero-engines under scenarios of working-condition shift. Reliab Eng Syst Saf2023; 237: 109350.
31.
ZhuKHuangCLiS, et al. Physics-informed Gaussian process for tool wear prediction. ISA Trans2023; 143: 548–556.
32.
WangJLiYZhaoR, et al. Physics guided neural network for machining tool wear prediction. J Manuf Syst2020; 57: 298–310.
33.
ShiJRiveraAWuD.Battery health management using physics-informed machine learning: online degradation modeling and remaining useful life prediction. Mech Syst Signal Process2022; 179: 109347.
34.
QiangBShiKLiuN, et al. Integrating physics-informed recurrent Gaussian process regression into instance transfer for predicting tool wear in milling process. J Manuf Syst2023; 68: 42–55.
35.
LiYWangJHuangZ.Physics-informed meta learning for machining tool wear prediction. J Manuf Syst2022; 62: 17–27.
36.
LiTSiXPeiH, et al. Data-model interactive prognosis for multi-sensor monitored stochastic degrading devices. Mech Syst Signal Process2022; 167: 108526.
37.
RakthanmanonTCampanaBMueenA, et al. Searching and mining trillions of time series subsequences under dynamic time warping. In: Proceedings of the 18th ACM SIGKDD international conference on knowledge discovery and data mining, Beijing, China, 19–23 August 2018, pp.262–270. ACM.
38.
Akhavan NiakiFMichelMMearsL. State of health monitoring in machining: extended kalman filter for tool wear assessment in turning of IN718 hard-to-machine alloy. J Manuf Process2016; 24: 361–369.
39.
LiTPeiHSiX, et al. Prognosis for stochastic degrading systems with massive data: a data-model interactive perspective. Reliab Eng Syst Saf2023; 237: 109344.
40.
KimSKimILeeJ, et al. Knowledge integration into deep learning in dynamical systems: an overview and taxonomy. J Mech Sci Technol2021; 35: 1331–1342.
41.
LuoMChangXNieL, et al. An adaptive semisupervised feature analysis for video semantic recognition. IEEE Trans Cybern2018; 48: 648–660.
42.
WillardJJiaXXuS, et al. Integrating scientific knowledge with machine learning for engineering and environmental systems. ACM Comput Surv2022; 55: 66.
43.
HuangQQianCLiC, et al. Tool remaining useful life prediction method based on multi-sensor fusion under variable working conditions. Machines2022; 10: 884.
44.
WangPLiuJQiJ, et al. Multi-task dual-level adversarial transfer learning boosted RUL estimation of CNC milling tools. Knowl-Based Syst2025; 312: 113152.
45.
ISO 8688-2:1989. Tool life testing in milling; part 2: end milling. Geneva: ISO.
46.
BleicherFRamsauerCMOswaldR, et al. Method for determining edge chipping in milling based on tool holder vibration measurements. CIRP Ann2020; 69: 101–104.
47.
LiuYXuRYangZ, et al. Effect of vibration frequency on tribological characteristics of vibration-assisted micro-nano machining. Tribol Int2025; 202: 110329.
48.
WuCFJ. On the convergence properties of the EM algorithm. Ann Stat1983; 11: 95–103.
49.
DempsterAPLairdNMRubinDB.Maximum likelihood from incomplete data via the EM algorithm. J R Stat Soc Ser B Stat Methodol1977; 39: 1–22.
50.
RathoreMSHarshaSP.An attention-based stacked BiLSTM framework for predicting remaining useful life of rolling bearings. Appl Soft Comput2022; 131: 109765.
