射出成型製程中機台校正效應對實驗設計優化法的效能影響之研究

The effect of machine calibration on the efficiency of Design of Experimental (DOE) method in injection molding process

許睿庭(Rui-Ting Xu)

指導教授 ： 林國賡

共同指導教授 ： 黃招財(Chao-Tsai Huang)

淡江大學/工學院/化學工程與材料工程學系碩士班/碩士(2020年)

https://doi.org/10.6846/TKU.2020.00826

摘要

近年來許多自動化技術被應用於在射出成型系統中，以實現工業4.0自動化。然而許多產品不斷地往輕、薄、短、小的趨勢發展，倘若在執行自動生產之前，沒有做好產品設計和模具開發上的管理，起初產品不當的設計將導致最後開模的過程中許多問題，為了更有效率克服上述所面臨的問題，以及如何保持產品良好的品質，因此目前產學界廣泛利用電腦輔助工程技術(computer-aided engineering, CAE)來協助；但是 CAE模擬分析與實際製造業的結果有所落差，即使從材料選定、模具尺寸到操作條件完成上述檢查模擬與實驗都相同的情況下，還是無法解決CAE模擬分析與實際實驗結果上的差異。有鑑於此，本研究第一部份為延續我們團隊之前所進行機台校正之方法，選定圓平板系統，持續探討應用CAE模擬分析與實驗結果之差異來探索機台之校正效能，再利用CAE模擬分析結合田口方法，探討機台校正效應對於田口法效能影響作基礎研究，結果顯示在相同射速50%設定情況下，實驗與模擬的收縮量值平均差距0.18 mm，但兩者內在驅動力一致；其中造成差異的來源可由壓力歷程曲線進行探索，藉此可進一步透過此等射出壓力歷程曲線校正後，可以有效改善模擬與實驗偏差值之差距。另外，在CAE結合DOE研究的部分，我們比較兩種最佳化選擇操作參數方法，兩者結果在未考慮機台校正前，模擬與實驗之間的偏差量值皆為0.12 mm，但經考慮機台校正後，模擬和實驗之間的偏差量值減小到0.05mm，準確性提高了58％；再則，第二部分的研究中，我們將探討不同機台校正效應，以驗證第一部分校正流程之可行性，特別著重探討在不同機台校正效應對於田口優化法效能影響之研究，經研究發現有時須透過射出成品校正可有效改善模擬與實驗偏差值之差距；另外，在CAE-DOE的部分，未考慮機台校正前，模擬與實驗之間的偏差量值為0.13 mm，考慮機台校正後，模擬和實驗之間的偏差量值減小到0.02 mm，準確性提高了85％；再來，第三部分研究中，我們嘗試利用CAE模擬分析結合反應曲面法，持續探討機台校正之效應如何影響反應曲面優化法在射出成型成品與製程之效能，結果顯示在CAE-RSM的部分，未考慮機台校正前，模擬與實驗之間的偏差量值為0.13 mm，但經考慮機台校正後，模擬和實驗之間的偏差量值減小到0.01 mm，準確性提高了92％。整體研究結果顯示機台校正效應對於在射出成型的產品開發相當的重要性。

關鍵字

射出成型 ； CAE模擬分析 ； 翹曲變形 ； 田口法 ； 反應曲面法

並列摘要

In recent years, many automation technologies have been used in injection molding systems to achieve Industry 4.0 automation. Before the automatic production is performed, if the product design and mold development management do not handle well, the product design and development will en encounter many problems. In order to solve those problems, computer-aided engineering (CAE) is widely used in the industry and academia to assist. However, it is very often to see that the CAE simulation result is not the same with the actual manufacturing obtained. Even all the conditions are the same for both simulation and experiment, the deviation is still there.
To solve this challenge, in this study the major content has been divided into three parts. In the first part, we have conducted the method of the machine calibration which was carried out by our team earlier
by selecting the circle plate as the system continuously. Using this system, the procedure of the machine calibration has been verified successfully. Specifically, using CAE-DOE method under the same injection velocity of 50% setting, the shrinkage difference between simulation and experiment is 0.18 mm.
Moreover, to discover the mechanism for both simulation and experiment, the injection pressure history curve has been applied as one pressure sensor installed in the entrance of gate for the mold. Using the concept of injection pressure history curve, the machine calibration can be performed smoothly. In addition, two operation parameter optimization methods have been proposed for CAE-DOE. Results showed that before doing the machine calibration, the deviation between the simulation and the experiment is 0.12 mm; After done the machine calibration, the deviation between the simulation and the experiment is reduced to 0.05 mm, and the accuracy is improved 58%. In the second part of this study, the machine calibration procedures have been performed in the different type of injection machine. We found that some calibration step needs to be modified, especially one more step based on the diameter of injected disc is added. Using this updated calibration procedures, in the CAE-DOE study, before the doing the machine calibration, the deviation between the simulation and the experiment is 0.13 mm; furthermore, after done the machine calibration, the deviation between the simulation and the experiment has been reduced to 0.02 mm.It shows the accuracy has been improved by 85%. Moreover, in the third part of this study, we have tried to use CAE simulation analysis combined with response surface method (CAE-RSM) to optimize the injection molding processing. Specifically, the effect of machine calibration on the performance of CAE-RSM has been considered. Specifically, before doing the machine calibration, the deviation between the simulation and the experiment is 0.13 mm; moreover, after the done the machine calibration, the deviation between the simulation and the experiment has been reduced to 0.01mm. The accuracy has been improved by 92%. Overall, it is quite important by considering the machine calibration effect in the quality control for injection molding production development.

並列關鍵字

Injection molding ； CAE simulation analysis ； Warpage ； Design of experiments ； Response surface method

參考文獻

[1] B. Ozcelik, and I. Sonat, (2009), "Warpage and structural analysis of thin shell plastic in the plastic injection molding", Materials & Design, 30(2), 367-375.

[2] 張榮語, (1995), 射出成型模具設計 : 材料特性 (初版. ed.). 臺北市: 高立圖書公司.

[3] 張元榕, 徐志忠, and許嘉翔, (2009), Moldex3D/Solid真實三維模流分析理論與應用,科盛科技股份有限公司.

[4] Y. Kuo, and M. R. Kamal. (1976). The fluid mechanics and heat transfer of injection mold filling of thermoplastic materials. AIChE Journal, 22(4), 661-669.

[5] 科盛科技公司, (2007), Moldex 3D模流分析技術與應用 (初版 ed.). 臺北縣土城市,全華圖書公司.