抗腐蝕合金是一種具有高度抗腐蝕能力的材料，近年來被廣泛應用於工業界中，主要是因為它卓越的抗腐蝕性可以降低由於腐蝕所造成的維修成本、破壞。為了設計適用於各種嚴苛環境的抗腐蝕合金，研究影響合金抗腐蝕能力的關鍵因素成為非常重要的研究主題。先前的研究提出臨界點蝕溫度 (Critical pitting temperature) 和點蝕電位 (Pitting potential) 這兩個用於衡量抗腐蝕能力的指標與金屬組成的經驗式之間存在線性關係，該經驗方程式中並未考慮環境參數對抗腐蝕能力的影響，然而，許多研究指出環境參數（例如酸鹼值、氯離子濃度、溶液溫度）對於合金抗腐蝕能力有顯著的影響。因此，本研究將透過機器學習方法探討合金組成與環境因素對合金抗腐蝕能力的影響，為抗腐蝕材料的開發提供重要的資訊。機器學習相對於傳統方法具有較低的計算成本和較短的開發週期，並且具有強大的數據處理能力和高度準確的預測性能。本研究將合金的組成和環境因素作為特徵值利用機器學習模型進行訓練和測試分析預測合金的抗腐蝕能力，透過本次研究發現環境參數對於合金抗腐蝕能力的預測至關重要，使我們能夠更全面地瞭解金屬組成和環境因素對抗腐蝕能力的影響。;Corrosion resistant alloys (CRAs), a class of materials that is of great importance in many applications because of its high corrosion resistance. In order to design good CRAs for various applications, understanding the key factors that determine corrosion resistance is crucial. Previous work showed that the critical pitting temperature can be estimated by a composition-depend equation, measure of alloying for resistance to corrosion (MARC), whereas the pitting corrosion resistance be predicted by a composition-depend equation, pitting resistance equivalent (PREN). These two empirical equations only depend on the alloy composition. However, environmental parameters, believed to be critical factors, are not considered in both empirical equations. In order to obtain comprehensive understanding of effects of the environmental parameters on corrosion resistance, machine learning technique, which has powerful data processing capabilities, is utilized in this work. The influence of metal composition and environmental parameters on corrosion resistance are studied using machine learning. This study shows that environmental parameters play important roles in predicting corrosion resistance and provides valuable insights into developing good CRAs.
