A Comparison of Input Data Used to Represent Phase Transformations during the Quenching of a Large Nuclear Forging

Abstract

The present work explores the importance of model parameters and input variables when simulating the quenching of thick sectioned nuclear forgings. The modelling approach adopted uses values of specific heat capacity, containing latent heat release, to simulate cooling curves; rather than calculating transformation kinetics based upon a mathematical model. Termed the effective specific heat (Cpeff), two different methods were used to establish values: differential scanning calorimetry (DSC) and thermos dynamic predictive software. Values were then included in finite element (FE) models to simulate the characteristic cooling at the mid-wall position in a thick section forging and were validated against production thermocouple data. The investigation found that the formation of ferrite, bainite and martensite or lower bainite were all represented by the data established using DSC and critical formation temperatures were comparable with others in the literature. Conversely, values calculated using the thermodynamic software failed to represent ferrite formation and predicted different critical transformation temperatures for bainite. The simulated cooling curve that used the software predicted Cpeff data was comparable to the thermocouple data either side of the bainite transformation, however during the transformation the effects of latent heat on cooling rate were over predicting leading to disparities. The equivalent DSC cooling curves produced a near exact match.