Application of Statistical Learning Methods to a Data Set of Measured Heat Transfer Coefficients for Continuous Casting

Abstract The heat transfer coefficient (HTC) between water and air/water sprays and the strand surface is a major parameter to simulate solidification in continuous casting processes. The HTC can be calculated via various equations from literature. However, based on in-house measurements, the application of statistical models for predicting local HTCs has become possible. Several models have been set up using Python coding and compared regarding the error metrics (root mean squared error and mean absolute error).