A novel AC loss measurement method for HTS coils based on parameter identification

Abstract

In the presence of AC current or AC magnetic field, superconductors will suffer AC loss. The measurement of AC loss is of great significance for applied superconductivity field. There are main three different methods for AC loss measurement. Among them, electrical measurement method is widely researched by different laboratories. The existed electrical measurement method mainly has two shortcomings: difficult to be used in HTS coils with high inductance and only available for periodic current conditions (only obtains average value of AC loss for one or several current periods and cannot obtain AC loss at any instant). In this paper, a novel AC loss measurement method based on parameter identification is proposed to solve the second shortcoming (only available for periodic current conditions). This novel method measures AC loss through identifying HTS coils’ resistance and inductance at any moment, which is completely different from previous measurement methods. The theoretical basis of this novel method mainly contains three parts: equivalent circuit of HTS coils, instantaneous power identification from Poynting theorem and parameter identification based on recursive least squares with fixed forgetting factor (FFRLS). Compared with AC loss measurement results of previous electrical method (integral method), the validity of this novel measurement method is demonstrated. In theory, this novel method can measure AC loss for any current waveform (periodic and aperiodic current), which greatly widens measurement range of previous electrical method. The principle of this novel method and detailed experimental results are discussed and analyzed in this paper.

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

“…The analysis focuses on coil A and B in the paper [11] to investigate the impact of the forgetting factor on the accuracy of AC loss measurement results under various conditions. By comparing and analysing the errors of AC loss results obtained using the FFRLS method, the influence of the forgetting factor on the accuracy of AC loss measurement is evaluated.…”

“…Conversely, when the estimation error decreases, the variable forgetting factor approaches 1. To make a distinction, the VFFRLS with (10) as the forgetting factor is referred to as A-VFFRLS, and the VFFRLS with (11) as the forgetting factor is referred to as B-VFFRLS.…”

“…The conventional least squares method considers all the previous data, which is not applicable for fast-changing system. To address this issue, fixed forgetting factor least squares method is employed for parameter identification of superconducting coils [11]. This method assigns greater weight to recent data, allowing for more accurate estimation of the dynamic system.…”

“…errors of the AC loss results obtained by FFRLS, A-VFFRLS, and B-VFFRLS relative to those of the integral method are compared and discussed. The forgetting factor of FFRLS is set to 0.9 for parameter identification [11], the pending constants of A-VFFRLS are selected in the optimal range, the sensitivity factor is set to 2000 and the minimum value of forgetting factor is set to 0.75 according to table 2.…”

“…To address these limitations, a method based on the recursive least squares method with fixed forgetting factor (FFRLS) is introduced for AC loss measurement [11]. This method identifies the parameters of the equivalent model using the current and voltage of the superconductor and calculates the AC loss via Joule's law.…”