鋁鎂合金材料銲接雖然有許多優異機械性質,但在氬銲銲接上其可銲性之條件範圍狹窄會有接合介面和融熔銲道形成硬而脆的介金屬化合物的困難點,一般對於銲接參數設定並沒有公式可循,完全憑藉專家過去的知識和經驗來設定,一旦超出專家經驗範圍,便無法有效設定最佳參數,本研究將發展一套解決鋁鎂合金銲接板參數多重品質特性實驗設計問題,探討銲接非破壞性品質特性─銲道厚度、銲道寬度、深寬比以及破壞性品質特性─拉伸、衝擊值等五個銲接品質特性,再應用理想解類似度順序偏好法 (Technique for Order Preference by Similarity to Ideal Solution) 與倒傳遞類神經網路 (Artificial Neural Network) 搜尋最佳架構,再結完全排列組合法 (All Combinations)找出鋁合金板材銲接參數最佳化。研究結果可提供銲接相關業者改善銲接效率。
The welding of aluminum magnesium alloy has superior mechanical characteristics, but the feasible setting for the welding parameters of the TIG (Tungsten Inter Gas Arc Welding) or GTAW (Gas Tungsten Arc Welding) have many difficulties due to some hard and crisp inter-metallic compounds created within the welding line. Normally, the setting of welding parameters does not have a formula to follow; it usually depends on experts' past knowledge and experiences. Once exceeding the rule of thumb, it becomes to be impossible to set up feasibly the optimal parameters. Consequently, this study will develop a solution to solve the problem of multiple quality characteristics when we weld the aluminum magnesium alloy plates. The select welding quality characteristics are non-destructive testing (Welding thickness, Welding width and Depth-width ratio) and destructive testing (Impact Test and Tensile Test). This paper uses TOPSIS (Technique for Order Preference by Similarity to Ideal Solution), ANN (Artificial Neural Network) and all combinations to find the optimal function framework of parameter design for welding of aluminum magnesium alloy. The research results can improve the welding efficiency for relevant welding industries.