Adaptive time stepping in quench simulation for superconducting magnets

Abstract

As an important issue of the quench simulation in large-scale applications, the effect of time step control is discussed on the well-known thermo-hydraulic model of superconducting magnets. Through our approach of model reduction, the reason of numerical anomaly is identified in terms of the time stepping around dynamical events, and just leads an effective countermeasure against the observed issue. Apart from the practical means out of such an investigation, a novel idea of non-iterative time step adaptation is introduced as a breakthrough not only being applied to the quench propagation model itself, but also to be extended to the common pitfalls of linearization scheme in general transient PDE (partial differential equation) problems.