Accurate prediction of the critical transformation temperatures \(Ac_{1}\) and \(Ac_{3}\) is one of the essential factors in the heat treatment of CA6NM cast martensitic stainless steel. In this study, the influence of alloy composition, heating rate, and prior austenite grain size on the critical temperatures was quantified using dilatometric measurements and statistical modeling. Six heat treatment conditions produced prior austenite grain sizes ranging from ~75 to 247 µm. Experimental results showed that grain size variations produced only minor changes in \(Ac_{1}\) and \(Ac_{3}\) , indicating a weak dependence of transformation temperatures on prior austenite grain size within the investigated range. In contrast, chemical composition and heating rate had strong effects, with \(Ac_{1}\) and \(Ac_{3}\) increasing by approximately 100 °C and 115 °C, respectively, as the heating rate increased from 0.1 to 10 °C s−1. Regression-based predictive model incorporating alloy composition and heating rate was developed and validated. The \(Ac_{1}\) model demonstrated strong agreement with experimental measurements (R2 = 0.972 and RMSE = 4.78 °C) and significantly improved prediction accuracy compared with thermodynamic calculations using Thermo-Calc and previously published empirical models. The results provide an improved framework for predicting transformation temperatures in CA6NM, under continuous heating conditions.
