自動化加工特徵辨識與工時預測的整合應用、驗證與學習

Integrated Application, Verification and Learning of Processing Feature Automatic Recognition and Processing Time Prediction

傅敬詒(JING-YI FU)

指導教授 ： 鍾文仁

中原大學/工學院/機械工程學系/碩士(2022年)

https://doi.org/10.6840/cycu202201414

摘要

近年來隨著智慧製造的蓬勃發展，傳統加工逐漸邁向自動化，模具產業亦不例外，面對多變複雜的產品需求，對於模具生產效率及精度的要求也日益提升，然而現今零件製程規劃大多仰賴具豐富經驗的技術人員，針對零件所需的製造工法進行工序的規劃，再使用電腦輔助製造(Computer-aided Manufacturing, CAM)軟體模擬加工時間，以人工來操作除了需要大量人事成本，繁瑣的操作流程也會增加製造所需的前置時間，甚至容易出現人為誤差而產生額外的時間成本，從而造成交期延誤；另一方面，產線自動化的趨勢使得工時預估的準確性更為重要，因此本研究開發一套引導系統，期望透過簡化操作流程來縮短加工前置作業時間，並結合迴歸分析來提高工時預估的準確度。
本研究以ASP.NET MVC Framework為架構，使用業界常用之電腦輔助設計(Computer-aided Design, CAD)軟體Siemens NX (又稱NX Unigraphics，或簡稱為UG)，並基於其二次開發模組建立網路化之線切割(Wire Electrical Discharge Machining, WEDM)、放電(Electrical Discharge Machining, EDM)加工模擬引導系統，並整合已有的CNC銑床加工引導流程，首先透過自動特徵辨識來識別出零件的WEDM及EDM加工特徵，再以自動化之加工設定搭配簡單的參數輸入引導使用者進行數控(Numerical Control, NC)編程，最後經由引導系統自動生成NC Code，EDM的部分則會在特徵辨識完成後，自動判斷加工範圍並產生電極頭。流程中取得之加工特徵參數將儲存於Microsoft SQL Server Express資料庫，並傳送至線性迴歸分析模型進行訓練，以實際加工時間做為訓練標籤，優化迴歸分析模型參數，以提高預估工時的準確性，使用本研究開發之引導系統能夠有效縮短84%的加工模擬操作時間，同時能夠快速地產生大量訓練資料供迴歸分析訓練使用，案例結果顯示，經迴歸優化後能夠將預估工時與實際工時的誤差最大值由約45%減少至約23%，達到生產效率的提升並提高預估工時準確度。

關鍵字

特徵辨識 ； CAM ； 二次開發 ； 工時預估 ； 機器學習

並列摘要

Recently with the rapid development of smart manufacturing, traditional processing is gradually moving towards automation, and the mold industry is no exception. Facing the ever-changing and complex product demands, the requirements for mold production efficiency and precision are also increasing. However, currently most of the parts process planning rely on experienced personnel. They will decide the order according to the manufacturing method for the parts, and then use Computer-aided Manufacturing (CAM) software to simulate the processing time. Manual operation not only requires a lot of human-resource costs but also increase the lead time required for manufacturing, and even because of those human errors, resulting in delays for delivery. On the other hand, the trend of production line automation makes the accuracy of processing time estimation more important. Therefore, this study develops a navigating system to shorten the processing lead time by simplifying the operation process, and combine regression analysis to improve the accuracy of processing time estimation.
This study uses ASP.NET MVC Framework as the framework, and applied Computer-aided Design (CAD) software, Siemens NX (also known as NX Unigraphics, or UG for short), to create a machining simulation navigating system on its secondary development module for Networked Wire Electrical Discharge Machining (WEDM), Electrical Discharge Machining (EDM), also integrates the existing CNC milling machine machining navigating process. This study identifies the WEDM of the part and EDM processing features through automatic feature recognition, and then navigate the user to perform Numerical Control (NC) programming with automatic processing settings and simple parameter input, and finally generate NC Code automatically through the system. For the EDM, the system will automatically decide the processing range and generate electrode tips after the feature identification is completed. The processing characteristic parameters will be stored in the Microsoft SQL Server Express database and sent to the linear regression analysis model for training. The actual processing time is used as the training label to optimize the parameters of the regression analysis model in order to improve the accuracy of the estimated working hours. The navigating system in this study can effectively shorten the processing simulation operation time by 84%, and at the same time a large amount of training data for regression analysis training can be generated. The errors of estimated machining time and real machining time is reduced from 45% to 23%, which improves production efficiency and the accuracy of estimated working hours.

並列關鍵字

feature recognition ； CAM ； secondary development ； processing time estimation ； machine learning

參考文獻

[1]. Omid Fatahi Valilai and Mahmoud Houshmand, 2010, "INFELT STEP: An Integrated and Interoperable Platform for Collaborative CAD/CAPP/CAM/CNC Machining Systems Based on STEP Standard", International Journal of Computer Integrated Manufacturing, 23(12), pp. 1095-1117.

[2]. A. K. Matta, D. R. Raju, and K. N. S. Suman, 2015, "The Integration of CAD/CAM and Rapid Prototyping in Product Development: A Review", Mater. Today, 2(4–5), pp. 3438-3445.

[3]. Jayakiran Reddy, E., Venkatachalapathi, N., and Pandu Rangadu, V., 2018, "Development of An Approach for Knowledge-based System for CAD Modelling", Mater. Today, 5(5-2), pp. 13375-13382.

[4]. Kong, L., Fuh, J. Y. H., Lee, K. S., Liu, X. L., Ling, L. S., Zhang, Y. F., and Nee, A.Y.C., 2003, "A Windows-native 3D Plastic Injection Mold Design System", Journal of Materials Processing Technology, 139(1-3), pp. 81-89.

[5]. Cui, H. Y., Tian, N., Li, R. C., and Men, X. H., 2014, "A Method of Parametric Design of Automobile Synchronizer Ring Based on UG Secondary Development Tools", Applied Mechanics & Materials, 716-717, pp. 635-638.