從射出壓力到光彈條紋：生成式模型於內部應力可視化之應用

射出成型製程中，材料內部應力雖深刻影響產品品質，卻難以於製程當下即時掌握。傳統光彈量測 多屬事後分析工具，不易直接應用於產線。本研究以模內壓力關鍵資訊為基礎，導入生成式模型建立製程 資料與光彈條紋之對應關係，使內部應力分布得以被預測與重建。本研究結果顯示，模型所生成的條紋在主要應力區域與整體分布趨勢上與實測結果具有良好一致性，且在有限資料條件下仍能維持穩定表現。 此方法有助於將內部應力評估由事後檢測前移至製程當下，為射出成型品質管理與智慧製造提供新的應用方向。

Internal residual stress in injection molding significantly affects product quality, yet remains difficult to monitor in real time during the manufacturing process. Conventional photoelastic measurement is primarily a post-process analysis tool and is not easily integrated into production lines. This study utilizes key in-mold pressure features and introduces a generative modeling approach to establish the relationship between process data and photoelastic fringe patterns, enabling the prediction and reconstruction of internal stress distributions. The results demonstrate that the generated fringe patterns show strong agreement with experimental measurements in both major stress regions and overall distribution trends, while maintaining stable performance even under limited data conditions. This approach supports shifting internal stress evaluation from post-production inspection to in-process assessment, providing a new direction for quality management and smart manufacturing in injection molding.