The Measurement of Heat Transfer Coefficients in Roll Casting

Two steady-state models of the twin-roll casting process and a transient model of the process have been developed. The models use the finite difference control volume method to solve the equations of heat flow. Experimentation has been done on a laboratory caster in conjunction with the modelling work. Steady-state Model I predicts the behaviour of the caster using assumed values of heat transfer coefficient along the length of contact between the strip and the roll. The model predictions of permissible casting speeds are for the most part close to those achieved in practice, but in some cases there are deviations between the predicted and actual speeds achievable. This is due to the uncertainty in the validity of the values of heat transfer coefficient used. The model can be used to simulate the effect of both internal cooling of the rolls and external spray-cooling of the rolls. Steady-state Model II uses a novel technique whereby measured values of roll subsurface temperature can be used to compute the values of heat transfer coefficient along the contact length. An instrumented roll was installed on the experimental caster to measure these temperatures, but problems arose due to the leakage of cooling water out to the roll surface during data acquisition trials and the resultant predictions of heat transfer coefficient are believed to be unrealistic. The transient model describes how the caster behaves under start-up conditions and how steady-state is approached with time. Its predictions are consistent with those of steady-state Model I using the same assumed values of heat transfer coefficient.