Analyzing dynamic resistance in high-temperature superconducting tapes by combining finite element method with machine learning

When an external alternating field reaches a threshold value, high-temperature superconductors (HTS) that are carrying direct current can exhibit dynamic resistance phenomenon. This phenomenon, often observed in tape applications, can be effectively studied using finite element methods (FEM). However, due to differences in production processes, HTS tapes have varying parameters, including magnetic-dependent critical current. This can pose a significant challenge when comparing dynamic resistance differences among HTS tapes. Due to the capability of machine learning to conveniently handle the nonlinear characteristics of superconductors and adapt to multivariate function fitting, this paper employs machine learning for fitting the critical current characteristics of tapes and applies it to calculate dynamic resistance in the FEM model. By employing machine learning to handle the critical current characteristics of various tapes, the FEM model showcases both feasibility and accuracy in the results.