機械手臂自動化加工有許多因素會影響加工成效，尤其參數設定更是機械手臂加工之關鍵問題，經由最佳化後的參數能夠大幅改善加工品質及生產效能，因此，本研究初期對機械手臂焊道精磨進行多目標參數最佳化，確保粗糙度及材料移除率，釐清參數變化對研磨結果影響。所採用方式為現今較為成熟的田口方法及直交表進行實驗設計，調整特定研磨參數，如研磨角度、側傾角、研磨速度等，將田口實驗所得出之結果作為依據，搭配灰色關聯分析法進行參數分析及綜合評比，完成多目標參數最佳化。研究後半段嘗試建立一套機械手臂自動化焊道研磨系統之焊道表面粗糙度與砂布輪磨耗監測系統，將先前得出之多目標最佳參數進行機械手臂研磨並收集訊號，通過機械手臂對平面工件上之焊道進行研磨過程中的機械手六軸扭矩訊號，工件的聲發射訊號及加速規振動訊號經由訊號分割，以及特徵運算等處理後再使用LSTM(Long Short-Term Memory）進行模型的訓練，以達到實時監測焊道表面粗糙度和砂布輪磨耗。研究成果顯示，經由參數最佳化之多目標最佳參數，在粗糙度及材料移除率的綜合表現上有優越的成果。而預測模型方面，焊道兩個方向的表面粗糙度預測分別可以達到8.692%、8.985%的平均絕對百分比誤差。砂布輪磨耗Rku值的預測可以達到4.19%的平均絕對百分比誤差。此外，在砂布輪磨耗等級分類能夠達到87.5%的準確率。

The purpose of this study is to optimize the robotic grinding process with flip discs and monitor the tool condition. To optimize the parameters of the weld bead grinding using robotic arm, Taguchi method and gray correlation analysis were employed to conduct the single-objective and multi-objective parameter optimization. Hereafter, this study attempted to develop a monitoring system to monitor the surface roughness of the ground weld bead and the wear of the flap discs during the grinding process. The vibration signals of accelerometer attached on the grinding machine, the torque signals of the robotic arm, and the acoustic emission signals from the workpiece were collected. Wavelet packet conversion (WPT), feature extraction, feature selection and long short-term memory (LSTM) technologies were utilized to develop the model to monitor the surface roughness of the weld bead and the wear of the flap discs. The used flip discs were examined by a digital microscope to measure the surface profile, and the root mean square (RΔq) was used as the labels to define the wear level of the flap discs. When predicting the RΔq value of the flap discs, it can reach 11.56% MAPE, and the prediction error of the roughness in two directions of the weld bead can reach 8.62% and 8.98% MAPE. The use of multiple sensors can obtain more processing-related information to improve the prediction performance of the model. Nevertheless, using only the acoustic emission signals can achieve the MAPE of 10.5% and 10.6%, which is not significantly lower than that with all the signals.

[1] V. Nguyen, F. Altarazi,T. Tran, "Optimization of Process Parameters for Laser Cutting Process of Stainless Steel 304: A Comparative Analysis and Estimation with Taguchi Method and Response Surface Methodology," Mathematical Problems in Engineering Volume 2022, 2022, Pages 14.
[2] F. Li ,Y. Xue ,Z. Zhang ,W. Song,J. Xiang , "Optimization of Grinding Parameters for the Workpiece Surface and Material Removal Rate in the Belt Grinding Process for Polishing and Deburring of 45 Steel," Mechanical Engineering Volume10, 2020,Pages 6314.
[3] M. I. Qazi ,R. Akhtar ,M. Abas ,Q. S. Khalid ,A. R. Babar ,C. I. Pruncu, " An Integrated Approach of GRA Coupled with Principal Component Analysis for Multi-Optimization of Shielded Metal Arc Welding (SMAW) Process," Materials Volume 13 ,2020, Pages 3457.
[4] S. Balaji, C. Sivakandhan , P. Maniarasan , D. Deepak , K. Senthamarai , S.V. Alagarsamy, " Effect of process parameters on machining behaviour using S/N ratio and ANOVA analysis," materialstoday:PROCEEDINGS Volume 74, Part 1, 2023, Pages 97-104
[5] P. Sivaiah , D. Chakradhar, " Modeling and optimization of sustainable manufacturing process in machining of 17-4 PH stainless steel," Measurement Volume 134, 2019, Pages 142-152 .
[6] M. Durairaj , D. Sudharsun , N. Swamynathan, " Analysis of Process Parameters in Wire EDM with Stainless Steel Using Single Objective Taguchi Method and Multi Objective Grey Relational Grade," Procedia Engineering Volume 64, 2013, Pages 868-877.
[7] I. Iglesias, M. A. Sebastián,J. E. Ares, "Overview of the state of robotic machining: Current situation and future potential," Procedia Engineering Volume 132, 2015, Pages 911-917.
[8] G. Singh , V. K. Banga , "Robots and its types for industrial applications," Materials Today: Proceedings Volume 60, Part 3, 2022, Pages 1779-1786
[9] Z. Pan,H. Zhang, "Robotic machining from programming to process control: a complete solution by force control," Industrial Robot, Volume35,2008,Pages 400-409.
[10] Y. Sun , D. J. Giblin , K. Kazerounian , "Accurate robotic belt grinding of workpieces with complex geometries using relative calibration techniques," Robotics and Computer-Integrated Manufacturing,Volume 25, 2009, Pages 204-210.
[11] Y.Lv, Z.Peng, C.Qu, D.Zhu , "An adaptive trajectory planning algorithm for robotic belt grinding of blade leading and trailing edges based on material removal profile model," Robotics and Computer-Integrated Manufacturing Volume 66, 2020,Pages 101987.
[12] G.Zhu , M. Cai , Z. Wang, L. Li , J. Zhao , B. Zhou , "A new calibration method for a dynamic coordinate system in a robotic blade grinding and polishing system based on the six-point limit principle," Robotics and Computer-Integrated Manufacturing Volume 83,2023, Pages 102561.
[13] C. Chen,Z. Cai, T. Chen, Z. Li, F. Yang, X. Liang, "A vision‑based calibration method for aero‑engine blade‑robotic grinding system," The International Journal of Advanced Manufacturing Techonology Volume 125 ,2023,Pages 2195-2209.
[14] H. Kang,J. Whan Noh,S. Jo Kwak ," Synchronization method for laser scanner and robot," International Conference on Control, Automation and Systems ( ICCAS) 2012,Pages 952-955.
[15] X. Xu, D. Zhu, H. Zhang, S. Yan , H. Ding, "TCP-based calibration in robot-assisted belt grinding of aero-engine blades using scanner measurements," The International Journal of Advanced Manufacturing Technology Volume90,2017, pages635–647.
[16] V. Pandiyan , T. Tjahjowidodo , "In-process endpoint detection of weld seam removal in robotic abrasive belt grinding process," The International Journal of Advanced Manufacturing Technology Volume 93, 2017,pages1699–1714 .
[17] Z. Wang, L. Zou, G. Luo, C. Lv, Y. Huang, "A novel selected force controlling method for improving robotic grinding accuracy of complex curved blade," ISA Transactions Volume 129,2022, Pages 642-658.
[18] D. Zhu , X. Xu, Z. Yang , K. Zhuang , S. Yan , H. Ding , Y. Huang, "Analysis and assessment of robotic belt grinding mechanisms by force modeling and force control experiments," Tribology International Volume 120, 2018, Pages 93-98.
[19] L. Zou, X. Liu, Y. Huang ,Y. Fei, "A numerical approach to predict the machined surface topography of abrasive belt flexible grinding," The International Journal of Advanced Manufacturing Technology Volume 104 , 2019, Pages2961-2970.
[20] X. Zhang , H. Chen , J. Xu , X. Song , J. Wang , X. Chen, "A novel sound-based belt condition monitoring method for robotic grinding using optimally pruned extreme learning machine," Journal of Materials Processing Technology Volume 260, 2018, Pages 9-19.
[21] N. Wang, G. P. Zhang, L. J. Ren, W. J. Pang, and Y. P. Wang, "Vision and sound fusion-based material removal rate monitoring for abrasive belt grinding using improved LightGBM algorithm," Journal of Manufacturing Processes Volume 66, 2021, Pages 281-292.
[22] V. Pandiyan, W. Caesarendra, T. Tjahjowidodo, and H. H. Tan, "In-process tool condition monitoring in compliant abrasive belt grinding process using support vector machine and genetic algorithm," Journal of Manufacturing Processes Volume 31, 2018, Pages 199-213.
[23] V. Pandiyan, S. Shevchik , K. Wasmer , S. Castagne , T. Tjahjowidodo, "Modelling and monitoring of abrasive finishing processes using artificial intelligence techniques: A review," Journal of Manufacturing Processes Volume 57, 2020, Pages 114-135
[24] J. V. Abellan-Nebot , F. R. Subiron, "A review of machining monitoring systems based on artificial intelligence process models," The International Journal of Advanced Manufacturing Technology ,Volume 47, 2010, pages237–257.
[25] M.A. Moore, "A review of two-body abrasive wear," Wear Volume 27, 1974, Pages 1-17
[26] V. Pandiyan, S. Shevchik , K. Wasmer , S. Castagne , T. Tjahjowidodo, "Modelling and monitoring of abrasive finishing processes using artificial intelligence techniques: A review," Journal of Manufacturing Processes Volume 57, 2020, Pages 114-135.
[27] S. Malkin ,C. Guo, " Grinding technology : theory and application of machining with abrasives" , 2nd ed ,New York: Industrial Press, 2008, Pages372.
[28] S. Hochreiter , J. Schmidhuber, "Long short-term memory,", Neural Comput, Volume 9, 1997,Pages 1735-1780.
[29] R. Kohavi , G. H. John, " Wrappers for feature subset selection,", Artificial Intelligence Volume 97, 1997, Pages 273-324.
[30] P. Domingos," A few useful things to know about machine learning," Communications of the ACM Volume 55, 2012,Pages 78–87.
[31] ABB,"ROBOTICS Product specification IRB 1200,"2019.
[32] A. K. Rouniyara, P. Shandilya," Multi-Objective Optimization using Taguchi and Grey Relational Analysis on Machining of Ti-6Al-4V Alloy by Powder Mixed EDM Process," materialstoday:PROCEEDINGS Volume 5, 2018, Pages 23779-23788.
[33] C. Feng , J. Kuo , T. Li Su," Optimization of multiple quality characteristics for polyether ether ketone injection molding process," Fibers and Polymers Volume 7,2006,Pages404–413.
[34] D. L. Fugal ,"Conceptual wavelets in digital signal processing: an in-depth, practical approach for the non-mathematician ,1st edition", San Diego," Calif.: Space & Signals Technical Pub, 2009, Pages 302 .
[35] A. Géron, "Hands-on machine learning with Scikit-Learn, Keras, and TensorFlow : concepts, tools, and techniques to build intelligent systems, 2nd edition. " ed. Beijing China ; Sebastopol, CA: O'Reilly Media, Inc, 2019, Pages 819.