Similar articles

Abstract: Accurate tool wear prediction is crucial for manufacturing efficiency, yet effectively using multi-domain sensor features is difficult due to redundant noise. There is a critical need to strategically leverage highly predictive strong features and potentially informative weak features. To address this issue, we propose CdualTAL, an improved Transformer-based encoder–attention–decoder algorithm. Its name represents the model’s key components: a correlation-adaptive feature selection algorithm module, a dual-channel Transformer encoder, an attention mechanism, and a long short-term memory (LSTM) decoder. CdualTAL employs a dual-channel encoder to independently process the full set of multi-domain features, along with a subset of strong features selected using a designed correlation-adaptive feature selection algorithm. A custom cross-attention mechanism is then used to fuse these representations, sharpening focus on strong features while judiciously integrating information from weak ones. Finally, a hierarchical LSTM decoder captures deep temporal dependencies. Validated on tool wear datasets, CdualTAL outperforms 11 state-of-the-art methods, achieving superior prediction stability and accuracy with an average R² of 0.983 and a root mean square error (RMSE) of 4.373.

CdualTAL：基于双通道Transformer和交叉注意力网络的多域刀具磨损预测

[1]Cai WL, Zhang WJ, Hu XF, et al., 2020. A hybrid information model based on long short-term memory network for tool condition monitoring. J Intell Manuf, 31(6):1497-1510.

[2]Chan YW, Kang TC, Yang CT, et al., 2022. Tool wear prediction using convolutional bidirectional LSTM networks. J Supercomput, 78(1):810-832.

[3]Chehrehzad M, Kecibas G, Besirova C, et al., 2024. Tool wear prediction through AI-assisted digital shadow using industrial edge device. J Manuf Process, 113:117-130.

[4]Chen XP, Wang R, Khalilian-Gourtani A, et al., 2024. A neural speech decoding framework leveraging deep learning and speech synthesis. Nat Mach Intell, 6(4):467-480.

[5]Duan J, Hu C, Zhan XB, et al., 2022. MS-SSPCANet: a powerful deep learning framework for tool wear prediction. Rob Comput Integr Manuf, 78:102391.

[6]Duan J, Zhang X, Shi TL, 2023. A hybrid attention-based paralleled deep learning model for tool wear prediction. Expert Syst Appl, 211:118548.

[7]Fawzi A, Balog M, Huang A, et al., 2022. Discovering faster matrix multiplication algorithms with reinforcement learning. Nature, 610(7930):47-53.

[8]Gao KP, Xu XX, Jiao SJ, 2022. Measurement and prediction of wear volume of the tool in nonlinear degradation process based on multi-sensor information fusion. Eng Fail Anal, 136:106164.

[9]Guo H, Lin X, Zhu KP, 2022. Pyramid LSTM network for tool condition monitoring. IEEE Trans Instrum Meas, 71:2509511.

[10]He ZP, Shi TL, Xuan JP, 2022. Milling tool wear prediction using multi-sensor feature fusion based on stacked sparse autoencoders. Measurement, 190:110719.

[11]Hou KL, Li RY, Liu XL, et al., 2025. Swin-fusion: an adaptive multi-source information fusion framework for enhanced tool wear monitoring. J Manuf Syst, 79:435-454.

[12]Huang ZW, Zhu JM, Lei JT, et al., 2020. Tool wear predicting based on multi-domain feature fusion by deep convolutional neural network in milling operations. J Intell Manuf, 31(4):953-966.

[13]Huang ZW, Shao JJ, Zhu JM, et al., 2024. Tool wear condition monitoring across machining processes based on feature transfer by deep adversarial domain confusion network. J Intell Manuf, 35(3):1079-1105.

[14]Kejriwal M, Kildebeck E, Steininger R, et al., 2024. Challenges, evaluation and opportunities for open-world learning. Nat Mach Intell, 6(6):580-588.

[15]Khoshouei M, Bagherpour R, Yari M, 2024. A smart look at monitoring while drilling (MWD) and optimizing using acoustic emission technique (AET). Sci Rep, 14(1):19766.

[16]Kumar AS, Agarwal A, Jansari VG, et al., 2025. Realizing on-machine tool wear monitoring through integration of vision-based system with CNC milling machine. J Manuf Syst, 78:283-293.

[17]Li WY, Fu HY, Han ZY, et al., 2022. Intelligent tool wear prediction based on Informer encoder and stacked bidirectional gated recurrent unit. Rob Comput Integr Manuf, 77:102368.

[18]Lu SF, Zhu YJ, Liu S, et al., 2022. A tool wear prediction model based on attention mechanism and IndRNN. Int Joint Conf on Neural Networks, p.1-7.

[19]Marani M, Zeinali M, Songmene V, et al., 2021. Tool wear prediction in high-speed turning of a steel alloy using long short-term memory modelling. Measurement, 177:109329.

[20]Max K, Kriener L, Pineda García G, et al., 2024. Learning efficient backprojections across cortical hierarchies in real time. Nat Mach Intell, 6(6):619-630.

[21]Mo CQ, Huang K, Ji HX, 2024. A lightweight and precision dual track 1D and 2D feature fusion convolutional network for machinery equipment fault diagnosis. Sci Rep, 14(1):31666.

[22]Oppliger J, Denner MM, Küspert J, et al., 2024. Weak signal extraction enabled by deep neural network denoising of diffraction data. Nat Mach Intell, 6(2):180-186.

[23]Ou JY, Liu HK, Huang GJ, et al., 2021. Tool wear recognition based on deep kernel autoencoder with multichannel signals fusion. IEEE Trans Instrum Meas, 70:3521909.

[24]Qiao HH, Wang TY, Wang P, et al., 2018. A time-distributed spatiotemporal feature learning method for machine health monitoring with multi-sensor time series. Sensors, 18(9):2932.

[25]Shah M, Vakharia V, Chaudhari R, et al., 2022. Tool wear prediction in face milling of stainless steel using singular generative adversarial network and LSTM deep learning models. Int J Adv Manuf Technol, 121(1-2):723-736.

[26]Shah M, Borade H, Dave V, et al., 2024. Utilizing TGAN and ConSinGAN for improved tool wear prediction: a comparative study with ED-LSTM, GRU, and CNN models. Electronics, 13(17):3484.

[27]Shi CM, Luo B, He SP, et al., 2020. Tool wear prediction via multidimensional stacked sparse autoencoders with feature fusion. IEEE Trans Ind Inform, 16(8):5150-5159.

[28]Si ZK, Si SM, Mu DQ, 2024. Efficient tool wear prediction in manufacturing: BiLPReS hybrid model with performer encoder. Arab J Sci Eng, 49(11):15193-15204.

[29]Turbé H, Bjelogrlic M, Lovis C, et al., 2023. Evaluation of post-hoc interpretability methods in time-series classification. Nat Mach Intell, 5(3):250-260.

[30]Wang CG, Wang GP, Wang T, et al., 2024. Exploring the processing paradigm of input data for end-to-end deep learning in tool condition monitoring. Sensors, 24(16):5300.

[31]Wang JJ, Yan JX, Li C, et al., 2019. Deep heterogeneous GRU model for predictive analytics in smart manufacturing: application to tool wear prediction. Comput Ind, 111:1-14.

[32]Wang S, Yu ZL, Guo YQ, et al., 2022. A CNN-LSTM-PSO tool wear prediction method based on multi-channel feature fusion. Mech Eng Sci, 4(2):39-48.

[33]Wu CH, Wu FZ, Qi T, et al., 2021. Fastformer: additive attention can be all you need.

[34]Yan BL, Zhu LD, Dun YC, 2021. Tool wear monitoring of TC4 titanium alloy milling process based on multi-channel signal and time-dependent properties by using deep learning. J Manuf Syst, 61:495-508.

[35]Yu XN, Wang HF, Wang JR, et al., 2024. A common feature-driven prediction model for multivariate time series data. Inform Sci, 677:120967.

[36]Zhao R, Wang JJ, Yan RQ, et al., 2016. Machine health monitoring with LSTM networks. 10^th Int Conf on Sensing Technology, p.1-6.