射出成型製程參數最佳化之研究: 結合資料包絡法分析與數學規劃

Tunghai University Institutional Repository > 工學院 > 工業工程與經營資訊學系所 > 碩博士論文 > Item 310901/24072

Please use this identifier to cite or link to this item: http://140.128.103.80:8080/handle/310901/24072

Title:	射出成型製程參數最佳化之研究: 結合資料包絡法分析與數學規劃
Other Titles:	Process Optimization for Plastic Injection Moldingby Data Envelope Analysis and Mathematical Programming
Authors:	黃振鈺 Huang,Chen-Yu
Contributors:	陳武林;黃欽印 Chen,Wu-Lin;Huang,Chin-Yin 工業工程與經營資訊學系
Keywords:	複迴歸分析;數學規劃;塑膠射出成型;資料包絡法;最佳化 Data Envelopment Analysis;Multiple Regression Analysis;Mathematical Programming;Injection Molding;Optimization
Date:	2013
Issue Date:	2014-02-19T07:03:33Z (UTC)
Abstract:	目前已發展出多種可以快速搜尋射出成型之參數組合的方法，但射出成型的參數設定依舊互相影響且組合眾多，持續的在尋找參數設定之間的關係，已經無法因應產品生命週期短且從中需要節省支出來創造利潤的時代了。過去對於塑膠射出成型探討的文獻中都針對單一或多個品質特性進行分析討論或採用雙反應曲面法進行兩個品質特性分析討論，文獻中未見有探討從成本角度搭配進行分析衡量。故此，本研究透過複迴歸分析找出品質函數關係，並求得三個品質參數(翹曲、收縮、頂出時體積收縮)之迴歸方程式，進而透過輸入製程參數便可獲得一組對應的翹曲、收縮及頂出時體積收縮比之輸出值，接著透過資料包絡法衡量出最有效率的參數組合，在經由成本方程式進行轉換，一樣透過複迴歸分析找出成本函數關係，並應用數學規劃LINGO求解一組最小成本參數組合，透過此組解與先前有效率之50組參數組合進行衡量，找出成本最小之預測值及製程參數值，最後透過Moldflow驗證最佳參數組合之準確性。本研究使用Moldflow軟體內建零件模型進行分析，以射出時間、射出壓力、保壓時間、保壓壓力、冷卻時間、冷卻劑溫度、熔劑溫度、模具溫度、開模時間等九個製程參數作為初始控制因子。首先，以Moldflow模擬軟體進行實際充填模擬，再進一步利用田口直交表及變異數分析找出製程之顯著因子（射出時間、射出壓力、保壓時間、保壓壓力），接著透過複迴歸分析結合資料包絡法搜尋出最佳化翹曲、收縮及頂出時體積收縮之最佳解組合。其參數組合透過成本方程式轉換成成本，接著透過複迴歸分析結合數學規劃搜尋出最小化成本參數組合，並與效率前緣線上之參數組合進行衡量，找出成本小且高效率之參數組合。故本研究提出一個可行的步驟方法，作為公司進行整體考量時的參考方法。 We developed a variety of injection molding can quickly search for a method of parameter combinations, but injection molding parameter settings still affect each other and the have many combinations. If we continued to looking for the relationship between the parameters settings, it can not be unable to the product of the life cycle and we need to save cost for create profits. In the past for the plastics injection molding of literature in general on a single quality characteristic analysis or use response surface method to do two quality characteristics analysis. So we consider the cost to dissus in this paper. In this thesis, use multiple regression analysis to identify multiple dependent variables for a functional relationship between one independent variable. And create a predictive model for warpage, shrinkage and volumetric shrinkage at ejection. Then use data envelopment analysis to find the optimum process parameter combinations. Finally, confirmed by the software Moldflow. And apply the mathematical programming method Lingo for solving a set of the minimisation of total cost parameter combinations, Through this group, an efficient solution with 50 sets of parameter combinations to measure, find the predictive value of the minimum cost and process parameters, and finally through the best parameter combination Moldflow verify the accuracy.　　In this study, use the software Moldflow of built part model to analysis. And nine process parameters are considered in the research: injection time, injection pressure, packing time, packing pressure, cooling time, cooling temperature, mold-open time, melt temperature, and mold temperature. First, use the software Moldflow to the actual filling simulation. And use Taguchi’s orthogonal array and ANOVA to find the significant factors of process. Finally, use combination of multiple regression analysis and data envelopment analysis to optimal the warpage, shrinkage and volumetric shrinkage at ejection. In this study, I anticipate that the procedure applied in my thesis will prove a useful modeling/optimization procedure that can be applied in practice to injection molding, and potentially to other polymer processes. Parameter combinations to transform cost that use the cost equation Use combination of multiple regression analysis and Mathematical Programming. In this study, I anticipate that the procedure applied in my thesis will prove a useful modeling/optimization procedure that can be applied in practice to injection molding, and potentially to other polymer processes.
Appears in Collections:	[工業工程與經營資訊學系所] 碩博士論文

Files in This Item:

File	Size	Format
101THU00030039-001.pdf	3323Kb	Adobe PDF	1395	View/Open

Loading...