Adaptive Mesh Refinement Method for Speeding Up Numerical Simulation of Electroslag Remelting Process

Abstract

The electromagnetic field and the droplet profile in electroslag remelting (ESR) process play a pivotal role in the quality of final steel products. Numerical simulation remains a central challenge to investigate ESR process with a high‐efficiency and high‐accuracy method. Herein, a strategy embedding an adaptive mesh refinement (AMR) method in 3D‐coupled model is proposed to speed up the simulation of ESR process. The coupling phenomena of the electromagnetic field and the droplet profile are successfully predicted using this method. Compared with the results of uniform mesh, the deviation of AMR results is less than 2.2%, whereas the computational speed of the AMR method is 10 times faster. Moreover, the effects of two key parameters in AMR method, maximum refinement level and refinement interval, on computational accuracy and speed are studied. With the increasing maximum refinement level, the computational accuracy is improved obviously first and then slightly, but the computation time increases exponentially. The increasing refinement interval has little effect on the computational accuracy but can greatly save the time of mesh operation. The proposed strategy can promote the application of numerical simulation in industrial ESR process.