Prediction of critical cooling rate under rapid heating rate and short holding time for pearlitic rail steel

Critical cooling rate ( V c ) is typically employed to assess the ability of forming martensite for pearlitic railway steels. Currently, the examination of V c is normally conducted by dilatometry analysis under a fully austenitizing condition. However, the martensite in railway steel is usually formed at the extremely fast heating rate and short holding time which corresponds to an insufficient austenitized state. This work focuses on the calculation of V c under insufficient austenitization state. A series of orthogonal experiments were designed to investigate the significance of heating rate ( v ), holding temperature ( T ), and holding time ( t ) which constitute different austenitization on V c and prior austenite grain size (PAGS). There is an extremely significant effect of T on V c and PAGS. An image recognition method based on machine learning models has been proposed to assist in calculating V c and PAGS. Moreover, a model to predict V c based on T and t was established for U71Mn pearlitic steel. This model exhibits a high concordance with the observed data and is hopeful to apply in other pearlitic railway steels.