- 學位論文
線切割放電加工之應用LSTM於斷線預測
Application of LSTM in Wire Cutting Electrical Discharge Machining in Wire Breakage Prediction
本文將於2029/08/27開放下載。若您希望在開放下載時收到通知，可將文章加入收藏。
摘要
放電線切割加工(Wire Electrical Discharge Machining, WEDM)是一種以非接觸式加工金屬的技術，採用放電所產生的高溫來熔解金屬，而此技術能夠用於切割各種複雜輪廓、內部結構及精密零件等，因此應用範圍也涵蓋了許多工業領域。但由於在放電加工的切割過程中，線電極會因長時間的放電磨損、參數的設置、加工零件的材料性質及表面平整性，導致其斷線的發生，進而使線切割機台的重新穿線，影響加工效率，且在斷線當下產生的火花，也會使加工零件的表面品質降低。 在線切割機放電切割工件時，機台的電流會隨著時間的加工而有變化，另參數的調整也是電流大小變化的因素，而線電極斷裂的當下，會伴隨著大量的能量，藉此觀察其斷線前的電流趨勢。因此，本研究為了預測線電極斷線的時間，利用徠通的AP-6040A線切割機進行實驗，將機台控制器的參數資料及機台的電流數據傳輸至資料庫。研究中，使用0.25mm的黃銅絲電極，於10mm厚度的SKD11合金鋼進行切割的動作，並調整適當的參數，藉由多次的加工過程蒐集加工時的電流數據，找出斷線前的電流趨勢變化，之後利用長短期記憶模型(Long Short-Term Memory, LSTM)提早10秒預測放電加工斷線的時間點，並使預測準確度高於94%。
並列摘要
Wire Electrical Discharge Machining (WEDM) is a non-contact metal machining technique that utilizes the high temperature generated by electric discharge to melt metal. This technology is capable of cutting various complex contours, internal structures, and precision parts, thus its application range also covers many industrial fields. However, during the cutting process of WEDM, wire electrodes will break due to long-term discharge wear due to prolonged electric discharge, parameter settings, material properties, and surface finish of the workpiece, leading to wire breakage and affecting machining efficiency. Moreover, sparks generated at the moment of wire breakage can also degrade the surface quality of the machined parts. During WEDM is processing, the machine's current varies with time and is influenced by parameter adjustments. The occurrence of wire electrode breakage is accompanied by a significant amount of energy, allowing for observation of the current trend just before breakage. Therefore, in this study, experiments were conducted using the AP-6040A WEDM machine, and parameters, along with real-time current data from the machine controller, were transmitted to a database. Using a 0.25mm brass wire electrode, cutting actions were performed on SKD11 alloy steel with thicknesses of 10mm. By collecting current data during multiple machining processes, the changing current trend before wire breakage was identified. Subsequently, a Long Short-Term Memory (LSTM) model was used to predict the time point of wire discharge machining breakage 10 seconds in advance, and the prediction accuracy was higher than 94%.
參考文獻
[1].Chen, Z., Zhou, H., Yan, Z., Han, F., and Yan, H., 2021, " A new high-speed observation system for evaluating the spark location in WEDM of Inconel 718", Journal of materials research and technology, 13, pp. 184-196.
[2].Pramanik, A., and Basak, A.K., 2018, " Sustainability in wire electrical discharge machining of titanium alloy: Understanding wire rupture", Journal of Cleaner Production, 198, pp. 472-479.
[3].Cabanes, I., Portillo, V., Marcos, E., and Sanchez, J.A., 2008, " On-line prevention of wire breakage in wire electro-discharge machining", Robotics and Computer-Integrated Manufacturing, 24, pp. 287-298.
[4].Chou, P.H., Huang, Y.R., and Yan , B.H., 2022, " The Study of Machine Learning for Wire RupturePrediction in WEDM manufacturing", Department of Mechanical Engineering, 119(1-2), pp. 1301-1311.
[5].Mather, I., Sarhan, A.A.D., Barzani, M.M., and Hambi, M., 2015, " Increasing the productivity of the wire-cut electrical discharge machine associated with sustainable production", Journal of Cleaner Production, 108, pp. 247-255.