Comparative study of recursive least squares with variable forgetting factor applied in AC loss measurement

AC loss has significant impact on the design and safe operation of superconducting power equipment. While there have been numerous simulation studies on AC loss, experimental measurement has proven challenging due to its relatively small magnitude compared to reactive power. In our previous work, a method based on fixed forgetting factor recursive least squares for measuring instantaneous AC loss is proposed. However, its applicability is limited due to the varying characteristics of each parameter in superconductors. This paper introduces four recursive least squares with variable forgetting factor algorithms and measures AC loss in ten different superconducting coils. The accuracy of these algorithms is analysed and compared using the integral method as benchmark. The results demonstrate that the recursive least squares with leverage-based multiple adaptive forgetting factors offers the widest range for AC loss measurements.