Fast Solution of Induction Heating Problems by Structure-Preserving Nonlinear Model Order Reduction

A fast approach to the numerical solution of induction heating problems is proposed. The projection space is efficiently determined by numerically computing a few Volterra kernels of the solution to the problem. Numerical results show that the construction of the reduced nonlinear model is performed at a computational cost that is orders of magnitude less than that for the numerical integration of the full problem. The reduced order model solution then allows accurately reconstructing the whole space-time distribution of magnetic and temperature fields at negligible computational cost.