Neural networks for solving the inverse heat transfer problem of continuous casting mould

Abstract

Accurate and real-time detection of mould transient heat transfer is a strong assurance for producing good initial solidification strand and accurately controlling the solidification process in continuous casting. Based on the measured data of mould temperatures during continuous casting of round billet, to decrease the calculation time for meeting the online monitoring and calculation requirement, the feasibility of using neural networks to solve the inverse problem of heat transfer for continuous casting mould was discussed. The results show that the neural network is faster for inverse model, and the calculation results by this method can correctly reflect the characteristics of non-uniform heat transfer, which provides a worthwhile and applicable method for online calculation and visual technology of heat transfer and solidification inside continuous casting mould.