射出成型是塑膠產品最為普遍的生產方式，但產品的缺陷卻要透過CAE模流分析或試模才能發現，因此衍伸出可製造性設計的概念，利用制定好的製造規範，快速篩檢CAD模型的設計問題。本研究便是將可製造性設計的概念程式化，對設計完成的CAD模型，進行可製造化分析，避免後期在模流分析或試模時，重複的搭建網格與修模，影響產品開發效率。可製造化分析是透過特徵辨識的資料，將CAD模型分解成獨立特徵，進行規範的比對，因此良好的特徵辨識，是可製造化分析精準的關鍵。本研究利用已知的特徵辨識資料，進行資料結構的統整，針對CAD模型中佔比最高的肋特徵，進行辨識演算法的改良，提高肋特徵辨識資料的正確率。此外，開發各項特徵尺寸之計算演算法，套用設計規範進行檢查，將設計不良處呈現予設計者，且提供建議的修改範圍。本研究之可製造化分析，可作為模具開發前或CAE分析前的良好工具。;Injection molding is the most common method to produce plastic products, but the defects of a CAD model design cannot be found until CAE analysis or mold tried is implmented. Therefore, a concept of design for manufacturing (DFM), was developed to apply design guideline to quickly check the problems in the CAD model. This study is to develope a program for DFM analysis automatically, examine the design of a CAD model, and hopefully can avoid repeated CAE analysis and mold trial. In the proposed method, a feature recognition algorithm is employed to extract important feature from a CAD model. The dimensions of each feature are then evaluated. A DFM analysis is then implmented automatically based on a set of design guidelines for each type of feature. Deficits on the CAD models are finally displayed both in graphical and numerical formats. The operating procedures of the proposed program will be described and result of several CAD models will be discussed to demonstrate the feasibility of the proposed method.
