Condition-Robust Tool Remaining Useful Life Prediction under Highly Variable Milling Conditions

Abstract:

Accurate prediction of tool Remaining Useful Life (RUL) is essential for reliable and cost-effective milling, particularly when machining commercially pure titanium (CP-Ti), where tool wear is highly irregular. In industrial practice, continuously varying cutting conditions further complicate tool condition monitoring and life prediction. This paper proposes a vibration-based monitoring framework for RUL prediction under strongly variable milling conditions. A hybrid deep learning model based on CNN–BiLSTM is developed to capture the non-stationary relationship between vibration signals and tool degradation. Performance is compared between a spindle-mounted, non-invasive sensor and a tri-axial accelerometer mounted on the machine table, and the benefit of sensor fusion is assessed. Results show that spindle vibration correlates strongly with tool degradation and achieves predictive performance close to that of multi-sensor configurations, while requiring minimal instrumentation. The proposed approach remains robust under variations in both operating conditions and wear mechanisms, enabling reliable RUL estimation in non-stationary milling environments.

• Periodical: Defect and Diffusion Forum (Volume 450)
• Pages: 19-30
• DOI: https://doi.org/10.4028/p-xi5kxU
• Online since: April 2026
• Authors: Stefania Ferrisi*, Sebastian Lang, Mohamadreza Afrasiabi, Rosita Guido, Domenico Umbrello, Giuseppina Ambrogio, Markus Bambach
• Keywords: Deep Learning, Milling Process, Remaining Useful Life, Tool Wear

References

[1] Sarraf M, Rezvani Ghomi E, Alipour S, Ramakrishna S, Liana Sukiman N. A state-of-the-art review of the fabrication and characteristics of titanium and its alloys for biomedical applications. Biodes Manuf 2022;5:371–95.

DOI: 10.1007/s42242-021-00170-3

[2] Yang H, Luo S. Titanium products for everyday life by Panzhihua Story. Titanium for Consumer Applications, Elsevier; 2019, p.13–25.

DOI: 10.1016/B978-0-12-815820-3.00012-5

[3] Campos F de O, Araujo AC, Kapoor SG. Experimental Comparison of Micromilling Pure Titanium and Ti–6Al–4V. J Micro Nanomanuf 2019;7.

DOI: 10.1115/1.4043501

[4] Proud L, Tapoglou N, Wika KK, Taylor CM, Slatter T. Role of CO2 cooling strategies in managing tool wear during the shoulder milling of grade 2 commercially pure titanium. Wear 2023;524–525:204798.

DOI: 10.1016/j.wear.2023.204798

[5] Srinivasa Reddy B, Pradeep Kumar M, Ashwin Jawahar S. Numerical and experimental validation of CP-Ti (Grade 2) under cryogenic condition. Mater Today Proc 2024.

DOI: 10.1016/j.matpr.2024.01.046

[6] Siahsarani A, Alinaghizadeh A, Azarhoushang B, Bayat M, Bösinger R. Sustainable and efficient cooling in titanium milling for dental applications: A study on supercritical CO2 + MQL with focus on tool wear and surface topography. Wear 2025;572–573:206051.

DOI: 10.1016/j.wear.2025.206051

[7] Assis JOM, Lauro CH, Pereira RBD, Brandão LC, Arruda ÉM, Davim JP. A Chip Formation Study of the Micro-Cutting of Commercially Pure Titanium. Metals (Basel) 2024;14:851.

DOI: 10.3390/met14080851

[8] Kumar J, Khamba JS. An experimental study on ultrasonic machining of pure titanium using designed experiments. Journal of the Brazilian Society of Mechanical Sciences and Engineering 2008;30.

DOI: 10.1590/S1678-58782008000300008

[9] Khan A, Maity K. Statistical modelling and machinability assessment of commercially pure titanium (CP-Ti) grade II: An experimental investigation. Measurement 2019;137:664–72.

DOI: 10.1016/j.measurement.2019.02.018

[10] Yedurkar DP, Schlech T, Sause MGR. A Systematic Review on Smart and Predictive Maintenance in Tool Condition Monitoring. IEEE Access 2025;13:106246–86.

DOI: 10.1109/access.2025.3579204

[11] Mohanraj T, Kirubakaran ES, Dinesh Kumar M, Naren ML, Suganithi Dharshan P, Mohamed I. Review of advances in tool condition monitoring techniques in the milling process. Meas Sci Technol 2024;35:092002.

[12] Mohamed A, Hassan M, M'Saoubi R, Attia H. Tool Condition Monitoring for High-Performance Machining Systems—A Review. Sensors 2022;22:2206.

DOI: 10.3390/s22062206

[13] Assafo M, Langendoerfer P. Tool remaining useful life prediction using feature extraction and machine learning-based sensor fusion. Results in Engineering 2025;28:107297.

DOI: 10.1016/j.rineng.2025.107297

[14] Guo D, Liu Y, Sun L, Li G. A Novel Kolmogorov–Arnold Attention Allocation Network for Cutting Tool Remaining Useful Life Prediction. Applied Sciences 2025;15:11549.

DOI: 10.3390/app152111549

[15] Chen X, Cheng K. Cutting Tool Remaining Useful Life Prediction Using Multi-Sensor Data Fusion Through Graph Neural Networks and Transformers. Machines 2025;13:1027.

DOI: 10.3390/machines13111027

[16] Mohanraj T, Shankar S, Rajasekar R, Sakthivel NR, Pramanik A. Tool condition monitoring techniques in milling process — a review. Journal of Materials Research and Technology 2020;9:1032–42.

DOI: 10.1016/j.jmrt.2019.10.031

[17] Yaoguo M, Changfeng Y, Liang T, Minchao C, Junxue R, Ming L, et al. Research progress on intelligent monitoring of machining condition based on indirect method. Advanced Engineering Informatics 2025;67:103518.

DOI: 10.1016/j.aei.2025.103518

[18] Čuš F, Župerl U. Real-Time Cutting Tool Condition Monitoring in Milling. Strojniški Vestnik – Journal of Mechanical Engineering 2011;57:142–50.

DOI: 10.5545/sv-jme.2010.079

[19] Sobol' IM, Shukman BV. Random and quasirandom sequences: Numerical estimates of uniformity of distribution. Math Comput Model 1993;18:39–45.

DOI: 10.1016/0895-7177(93)90160-Z

[20] Hochreiter S, Schmidhuber J. Long Short-Term Memory. Neural Comput 1997;9:1735–80.

DOI: 10.1162/neco.1997.9.8.1735

[21] O'shea K, Nash R. An Introduction to Convolutional Neural Networks. n.d.

[22] Kiranyaz S, Avci O, Abdeljaber O, Ince T, Gabbouj M, Inman DJ. 1D convolutional neural networks and applications: A survey. Mech Syst Signal Process 2021;151:107398.

DOI: 10.1016/j.ymssp.2020.107398

[23] Paszke A, Gross S, Massa F, Lerer A, Bradbury J, Chanan G, et al. PyTorch: An Imperative Style, High-Performance Deep Learning Library 2019.