切削加工中，刀具狀態與加工效率及產品品質高度相關。因此，利用各式感測器於加工中進行刀具狀態監測（TCM）已獲廣泛研究；然而，現有文獻多聚焦於當前狀態之估計，較少對未來演變趨勢進行預測。此外，多數工業機台設計時並未預留感測器安裝空間，導致傳統架設方式程序繁瑣且易干擾實際加工。為克服上述挑戰，本研究針對車削製程，提出一套整合多感測器融合與 Transformer 架構之自適應刀具磨耗預測系統。該系統不僅能在加工中途同步輸出當前磨耗估計值與最終磨耗預測值，且在數據蒐集與訓練階段，無需於中途停機量測刀具，可在不中斷製程之前提下產生訓練所需之資料標籤。在硬體創新方面，本研究基於聽診器之聲音傳導原理，研發出一種創新的聲音感測技術，藉由機台外部的非侵入式麥克風配置，在不干擾加工的前提下精準獲取來自加工機台中的關鍵切削聲學訊號。本預測系統整合了主軸電流、刀把三軸振動及切削音訊等共 6 個通道訊號，經由時域與頻域特徵提取，並針對聲學訊號額外計算梅爾倒頻譜（MFCC）之統計特徵；經相關性篩選後，輸入至自建之 Transformer 編碼器模型進行刀具磨耗之提早預測。實驗結果顯示：在直線車削案例中，多感測器融合系統之最終磨耗預測平均絕對百分比誤差（Average MAPE）達 13.14%。若僅使用本研究提出之創新型聲音感測訊號，其預測誤差為 17.16%，顯著優於傳統麥克風架設法的 20.35%。在更複雜的輪廓曲面車削實驗中，多感測器系統之最終預測誤差仍可維持在 13.55%；而單獨使用創新聲學感測訊號的預測誤差為 14.41%，相較於傳統麥克風架設法的 19.40% 亦展現出顯著優勢。本研究不僅實現了刀具磨耗的預警與前瞻性控制，更證實了此創新的聲音感測方案在真實工業加工環境中的高度可行性與應用潛力。

In cutting processes, tool condition directly dictates machining efficiency and product quality. Although Tool Condition Monitoring (TCM) has been widely studied, most research focuses on estimating current states rather than predicting future trends. Moreover, traditional sensor installations are invasive and interfere with operations. To address these challenges, this study proposes an adaptive tool wear prediction system for turning processes using multi-sensor fusion with a Transformer-based architecture. The system provides real-time wear estimation and final wear prediction during turning operation, while requiring no intermediate machine downtime for tool wear measurement and training data collection. A key hardware innovation is a stethoscope-inspired, non-invasive acoustic sensing technique mounted outside the machine to capture critical cutting sounds. The system integrates six signal channels: spindle current, tri-axial tool holder vibration, and acoustic sound. After extracting time-, frequency-domain, and Mel-Frequency Cepstral Coefficient (MFCC) features, selected data are processed by a Transformer encoder for early prediction. Experimental results show that for straight turning, the multi-sensor system achieves a final wear prediction Average MAPE of 13.14%. Using only the novel acoustic signal yields 17.16% error, outperforming traditional microphone setups (20.35%). In complex contour turning, the multi-sensor error remains at 13.55%, while the novel acoustic signal alone achieves 14.41%, again significantly superior to traditional setups (19.40%). This study demonstrates a highly practical solution for proactive tool management in real industrial environments.

[1] F. W. Taylor, On the art of cutting metals (no. 1119). American society of mechanical engineers, 1907.
[2] D. Y. Pimenov, L. R. da Silva, A. Ercetin, O. Der, T. Mikolajczyk, and K. Giasin, "State-of-the-art review of applications of image processing techniques for tool condition monitoring on conventional machining processes," The International Journal of Advanced Manufacturing Technology, vol. 130, no. 1, pp. 57–85, 2024.
[3] H. Ravindra, Y. Srinivasa, and R. Krishnamurthy, "Acoustic emission for tool condition monitoring in metal cutting," Wear, vol. 212, no. 1, pp. 78–84, 1997.
[4] S. K. Choudhury and S. Rath, "In-process tool wear estimation in milling using cutting force model," Journal of Materials Processing Technology, vol. 99, no. 1-3, pp. 113–119, 2000.
[5] F. Alonso and D. Salgado, "Analysis of the structure of vibration signals for tool wear detection," Mechanical systems and signal processing, vol. 22, no. 3, pp. 735–748, 2008.
[6] A. Akbari, M. Danesh, and K. Khalili, "A method based on spindle motor current harmonic distortion measurements for tool wear monitoring," Journal of the Brazilian Society of Mechanical Sciences and Engineering, vol. 39, no. 12, pp. 5049–5055, 2017.
[7] Z. He, T. Shi, J. Xuan, and T. Li, "Research on tool wear prediction based on temperature signals and deep learning," Wear, vol. 478, p. 203902, 2021.
[8] Y. Zeng, R. Liu, and X. Liu, "A novel approach to tool condition monitoring based on multi-sensor data fusion imaging and an attention mechanism," Measurement Science and Technology, vol. 32, no. 5, p. 055601, 2021.
[9] D. H. Tien, N. V. Thien, T. T. T. Pham, and T. D. Nguyen, "Combined analysis of acoustic emission and vibration signals in monitoring tool wear, surface quality and chip formation when turning SCM440 steel using MQL," EUREKA: Physics and Engineering (2023),(1), pp. 86–101, 2023.
[10] M. S. Babu and T. B. Rao, "Multi-sensor heterogeneous data-based online tool health monitoring in milling of IN718 superalloy using OGM (1, N) model and SVM," Measurement, vol. 199, p. 111501, 2022.
[11] T. Mohanraj, S. Shankar, R. Rajasekar, N. Sakthivel, and A. Pramanik, "Tool condition monitoring techniques in milling process—A review," Journal of Materials Research and Technology, vol. 9, no. 1, pp. 1032–1042, 2020.
[12] G. Serin, B. Sener, A. M. Ozbayoglu, and H. O. Unver, "Review of tool condition monitoring in machining and opportunities for deep learning," The International Journal of Advanced Manufacturing Technology, vol. 109, no. 3, pp. 953–974, 2020.
[13] L. I. Guochao, S. Xinhang, S. U. N. Li, F. U. Bofeng, Y. Lei, and Z. Honggen, "Application of audible sound signals in tool wear monitoring: a review," Journal of Advanced Manufacturing Science and Technology, vol. 5, no. 1, p. 2025003, 2025, doi: 10.51393/j.jamst.2025003.
[14] F. Huda, K. Karjuni, and M. Rusli, "Cutting tool wear analysis using sound signal and simple microphone," in IOP Conference Series: Materials Science and Engineering, 2020, vol. 830, no. 4: IOP Publishing, p. 042028.
[15] Z. Li, X. Liu, A. Incecik, M. K. Gupta, G. M. Królczyk, and P. Gardoni, "A novel ensemble deep learning model for cutting tool wear monitoring using audio sensors," Journal of Manufacturing Processes, vol. 79, pp. 233–249, 2022.
[16] C. a. Zhou, K. Guo, and J. Sun, "Sound singularity analysis for milling tool condition monitoring towards sustainable manufacturing," Mechanical Systems and Signal Processing, vol. 157, p. 107738, 2021.
[17] Z. Li, R. Liu, and D. Wu, "Data-driven smart manufacturing: Tool wear monitoring with audio signals and machine learning," Journal of Manufacturing Processes, vol. 48, pp. 66–76, 2019.
[18] J. Steckel, A. Aerts, E. Verreycken, D. Laurijssen, and W. Daems, "Tool Wear Prediction in CNC Turning Operations using Ultrasonic Microphone Arrays and CNNs," ArXiv, vol. abs/2406.08957, 2024–06–13 2024, doi: 10.48550/arxiv.2406.08957.
[19] S. Ferrisi, R. Guido, D. Lofaro, G. Zangara, D. Conforti, and G. Ambrogio, "A CNN architecture for tool condition monitoring via contactless microphone: regression and classification approaches," The International Journal of Advanced Manufacturing Technology, vol. 136, no. 2, pp. 601–618, 2025/01/01 2025, doi: 10.1007/s00170-024-14860-6.
[20] A. Kothuru, S. Nooka, and R. Liu, "Audio-Based Tool Condition Monitoring in Milling of the Workpiece Material With the Hardness Variation Using Support Vector Machines and Convolutional Neural Networks," Journal of Manufacturing Science and Engineering, 2018–08–03 2018, doi: 10.1115/1.4040874.
[21] J. Kanakasuntharam and A. Wijethunge, "CNN and ANN Based Tool Condition Monitoring in Gear Machining Using Audio and Vibration Signals Via Cost Effective Sensors," 2024–03–05 2024, pp. 1–6, doi: 10.1109/esci59607.2024.10497391. [Online]. Available: https://consensus.app/papers/cnn-and-ann-based-tool-condition-monitoring-in-gear-kanakasuntharam-wijethunge/ad9307454c9f5eb496575d3f00c0798c/
[22] A. Kothuru, S. P. Nooka, and R. Liu, "Application of audible sound signals for tool wear monitoring using machine learning techniques in end milling," The International Journal of Advanced Manufacturing Technology, vol. 95, no. 9, pp. 3797–3808, 2018.
[23] C. Cooper, J. Zhang, R. X. Gao, P. Wang, and I. Ragai, "Anomaly detection in milling tools using acoustic signals and generative adversarial networks," Procedia Manufacturing, vol. 48, pp. 372–378, 2020.
[24] M.-K. Liu, Y.-H. Tseng, and M.-Q. Tran, "Tool wear monitoring and prediction based on sound signal," The International Journal of Advanced Manufacturing Technology, vol. 103, no. 9, pp. 3361–3373, 2019.
[25] H. Lv, J. Chen, T. Pan, T. Zhang, Y. Feng, and S. Liu, "Attention mechanism in intelligent fault diagnosis of machinery: A review of technique and application," Measurement, vol. 199, p. 111594, 2022.
[26] B. Guo, Q. Zhang, Q. Peng, J. Zhuang, F. Wu, and Q. Zhang, "Tool health monitoring and prediction via attention-based encoder-decoder with a multi-step mechanism," The International Journal of Advanced Manufacturing Technology, vol. 122, no. 2, pp. 685–695, 2022.
[27] W. Ma, L. Guo, H. Gao, Y. Yu, and M. Qian, "A health indicator construction method based on self-attention convolutional autoencoder for rotating machine performance assessment," Measurement, vol. 204, p. 112108, 2022.
[28] H. Liu, Z. Liu, W. Jia, X. Lin, and S. Zhang, "A novel transformer-based neural network model for tool wear estimation," Measurement Science and Technology, vol. 31, no. 6, p. 065106, 2020.
[29] C. Dong and J. Zhao, "An augmented autoencoder with multi-head attention for tool wear prediction in smart manufacturing," IEEE Access, vol. 12, pp. 79128–79137, 2024.
[30] Z. Si, S. Si, and D. Mu, "Efficient tool wear prediction in manufacturing: BiLPReS hybrid model with performer encoder," Arabian Journal for Science and Engineering, vol. 49, no. 11, pp. 15193–15204, 2024.
[31] J. Sun, H. Shi, J. Zhu, B. Song, Y. Tao, and S. Tan, "Self-attention-based Multi-block regression fusion Neural Network for quality-related process monitoring," Journal of the Taiwan Institute of Chemical Engineers, vol. 133, p. 104140, 2022.
[32] H. Zhang, W. Yang, W. Xu, C. Wu, J. Ma, J. Liu, and J. He, "Tool wear prediction with few samples based on stacked BiLSTM and parallel residuals with multi-head self-attention," Measurement, p. 119078, 2025.
[33] H. Wang, S. Wang, W. Sun, and J. Xiang, "Multi-sensor signal fusion for tool wear condition monitoring using denoising transformer auto-encoder Resnet," Journal of Manufacturing Processes, vol. 124, pp. 1054–1064, 2024.
[34] W. Jia, W. Wang, Z. Li, and H. Li, "Prediction of tool wear in sculpture surface by a new fusion method of temporal convolutional network and self-attention," The International Journal of Advanced Manufacturing Technology, vol. 121, no. 3, pp. 2565–2583, 2022.
[35] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, "Attention is all you need," Advances in neural information processing systems, vol. 30, 2017.
[36] ISO 3685:1993: Tool-life testing with single-point turning tools, International Organization for Standardization, Geneva, Switzerland, 1993. [Online]. Available: https://www.iso.org/standard/9151.html
[37] A. E. Diniz, Á. R. Machado, and J. G. Corrêa, "Tool wear mechanisms in the machining of steels and stainless steels," The International Journal of Advanced Manufacturing Technology, vol. 87, no. 9, pp. 3157–3168, 2016.
[38] Á. R. Machado and A. E. Diniz, "Tool wear analysis in the machining of hardened steels," The International Journal of Advanced Manufacturing Technology, vol. 92, no. 9, pp. 4095–4109, 2017.
[39] S. Dolinšek, B. Šuštaršič, and J. Kopač, "Wear mechanisms of cutting tools in high-speed cutting processes," Wear, vol. 250, no. 1-12, pp. 349–356, 2001.
[40] L. Equeter, F. Ducobu, and P. Dehombreux, "Cutting Tools Replacement: Toward a Holistic Framework," IFAC-PapersOnLine, vol. 53, no. 3, pp. 227–232, 2020.
[41] J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, "BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding," Minneapolis, Minnesota, June 2019: Association for Computational Linguistics, in Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pp. 4171–4186, doi: 10.18653/v1/N19-1423. [Online]. Available: https://aclanthology.org/N19-1423/
[42] P. Shaw, J. Uszkoreit, and A. Vaswani, "Self-attention with relative position representations," in Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers), 2018, pp. 464–468.
[43] C. Raffel, N. Shazeer, A. Roberts, K. Lee, S. Narang, M. Matena, Y. Zhou, W. Li, and P. J. Liu, "Exploring the limits of transfer learning with a unified text-to-text transformer," Journal of machine learning research, vol. 21, no. 140, pp. 1–67, 2020.
[44] A. Kazemnejad, I. Padhi, K. Natesan Ramamurthy, P. Das, and S. Reddy, "The impact of positional encoding on length generalization in transformers," Advances in Neural Information Processing Systems, vol. 36, pp. 24892–24928, 2023.
[45] R. Child, S. Gray, A. Radford, and I. Sutskever, "Generating long sequences with sparse transformers," arXiv preprint arXiv:1904.10509, 2019.
[46] C. Chung, B.-X. Cai, and B. Chinomona, "Prescriptive tool wear Control: Tool wear early prediction and cutting parameter regulation for discrete turning operations," Engineering Applications of Artificial Intelligence, vol. 156, p. 111238, 2025.
[47] Z. K. Abdul and A. Al-Talabani, "Mel Frequency Cepstral Coefficient and its Applications: A Review," IEEE Access, vol. 10, pp. 122136–122158, 2022, doi: 10.1109/access.2022.3223444.
[48] T. Benkedjouh, N. Zerhouni, and S. Rechak, "Tool condition monitoring based on mel-frequency cepstral coefficients and support vector regression," 2017–10–01 2017, pp. 1–5, doi: 10.1109/icee-b.2017.8192024. [Online]. Available: https://consensus.app/papers/tool-condition-monitoring-based-on-melfrequency-cepstral-benkedjouh-zerhouni/556b2f8f6eca529a8852b1c94f029ae7/
[49] T. K. Jan Hauke, "Comparison of values of Pearson’s and Spearman’s correlation coefficient on the same sets of data," QUAESTIONES GEOGRAPHICAE 30(2), 2011.
[50] Y. Zhang, K. Zhu, X. Duan, and S. Li, "Tool wear estimation and life prognostics in milling: Model extension and generalization," Mechanical Systems and Signal Processing, vol. 155, p. 107617, 2021.
[51] C. Yang, Y. Shi, H. Xin, T. Zhao, N. Zhang, and C. Xian, "Tool wear prediction model based on wear influence factor," The International Journal of Advanced Manufacturing Technology, vol. 129, no. 3, pp. 1829–1844, 2023.
[52] M. Geva, R. Schuster, J. Berant, and O. Levy, "Transformer feed-forward layers are key-value memories," in Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, 2021, pp. 5484–5495.
[53] Z. Chen, V. Badrinarayanan, C.-Y. Lee, and A. Rabinovich, "Gradnorm: Gradient normalization for adaptive loss balancing in deep multitask networks," in International conference on machine learning, 2018: PMLR, pp. 794–803.
[54] L. Liu, X. Liu, J. Gao, W. Chen, and J. Han, "Understanding the difficulty of training transformers," in Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2020, pp. 5747–5763.
[55] T. Yu, S. Kumar, A. Gupta, S. Levine, K. Hausman, and C. Finn, "Gradient surgery for multi-task learning," Advances in neural information processing systems, vol. 33, pp. 5824–5836, 2020.
[56] N. Wang and X. Zhao, "Enformer: Encoder-based sparse periodic self-attention time-series forecasting," IEEE Access, vol. 11, pp. 112004–112014, 2023.
[57] H. Yun, H. Kim, E. Kim, and M. B. Jun, "Development of internal sound sensor using stethoscope and its applications for machine monitoring," Procedia Manufacturing, vol. 48, pp. 1072–1078, 2020.
[58] H. Yun, H. Kim, Y. Jeong, and M. Jun, "Autoencoder-based anomaly detection of industrial robot arm using stethoscope based internal sound sensor," Journal of Intelligent Manufacturing, vol. 34, pp. 1427–1444, 2021–12–02 2021.
[59] E. Kim, D. Mun, M. B. Jun, and H. Yun, "Operation and productivity monitoring from sound signal of legacy pipe bending machine via convolutional neural network (CNN)," International Journal of Precision Engineering and Manufacturing, vol. 25, no. 7, pp. 1437–1456, 2024.
[60] E. Kim, H. Yun, O. C. Araujo, and M. B. Jun, "Sound recognition based on convolutional neural network for real-time cutting state monitoring of tube cutting machine," International Journal of Precision Engineering and Manufacturing-Smart Technology, vol. 1, no. 1, pp. 1–18, 2023.
[61] E. Kim, T. Bui, J. Yuan, S. C. Mouli, B. Ribeiro, R. A. Yeh, M. P. Fassnacht, and M. B. Jun, "Online real-time machining chatter sound detection using convolutional neural network by adopting expert knowledge," Manufacturing Letters, vol. 41, pp. 1386–1397, 2024.
[62] J. Lee, S. Akin, Y. Sim, H. Lee, E. Kim, J. Nam, K. Song, and M. B. Jun, "A stethoscope-guided interpretable deep learning framework for powder flow diagnosis in cold spray additive manufacturing," Manufacturing Letters, vol. 41, pp. 1515–1525, 2024.
[63] E. Kim, J. Jeon, Y. Kim, H. Yun, J. Wellman, Y. W. Choi, S. W. Lee, M. B.-G. Jun, and J. Lee, "Control-resilient roller wear prediction for thin wire flattening process via an internal sound-guided dynamic conditional network," International Journal of Precision Engineering and Manufacturing-Green Technology, vol. 12, no. 3, pp. 919–934, 2025.
[64] C. Han, Y. Sim, H. Jung, J. Lee, H. Lee, Y. S. Kang, S. Woo, G. Kim, H. W. Park, and M. B.-G. Jun, "IMPACT: Industrial Machine Perception via Acoustic Cognitive Transformer," arXiv preprint arXiv:2507.06481, 2025.
[65] C. Han, Y. S. Kang, Y. Sim, H. W. Park, and M. B.-G. Jun, "LISTEN: lightweight industrial sound-representable transformer for edge notification," Advanced Engineering Informatics, vol. 76, p. 104944, 2026.
[66] N. Ambhore, D. Kamble, S. Chinchanikar, and V. Wayal, "Tool condition monitoring system: A review," Materials Today: Proceedings, vol. 2, no. 4-5, pp. 3419–3428, 2015.
[67] M. Kuntoğlu, E. Salur, M. K. Gupta, M. Sarıkaya, and D. Y. Pimenov, "A state-of-the-art review on sensors and signal processing systems in mechanical machining processes," The International Journal of Advanced Manufacturing Technology, vol. 116, no. 9, pp. 2711–2735, 2021.
[68] B. Li, "A review of tool wear estimation using theoretical analysis and numerical simulation technologies," International Journal of Refractory Metals and Hard Materials, vol. 35, pp. 143–151, 2012.
[69] M. Bhuiyan and I. Choudhury, "Review of sensor applications in tool condition monitoring in machining. Comprehensive Mater Process 13: 539–569," ed, 2014.