在鞋業中底發泡製程中,所使用之高分子原料的種類繁多,故針對不同材料特性之掌握不易。在實際製程上也因此普遍存在產品尺寸誤差與製程設定條件變異大的問題。本研究主要目的係在射出發泡成型系統上建立一個塑料升溫歷程的量測方法,進一步將其量測數據結合由實驗取得的同類型發泡材料之熱傳導係數,兩者數據代入類神經訓練法後,並取得較佳之溫度變化的熱傳導係數。最終將較佳之材料參數導入實際鞋型模具,使有限元素法進行鞋型塑料升溫歷程之分析。取得發泡鞋型中底塑料溫度分佈之數據,將有助於塑料成型尺寸的估算與發泡成品品質之提升。 Various types of polymer materials are used in manufacture of midsoles in the footwear industry. The variety of those materials as well as their diverse properties pose difficulties in the midsoles manufacturing process. This often causes an instability in manufacturing configurations and a large tolerance in molding sizes. In this study, we create and implement a material temperature tracking method for a foam molding system used in the shoe manufacturing process. A model of temperature-to-thermal-conductivity-coefficient mapping is obtained by using a self-trained neural network. This model is then used by Finite Element Method to carry out analysis of heat distribution. This improves the accuracy in the estimation of the molding size and helps with the quality of the final products.