Intelligent fault diagnosis and protection in DG-connected systems using resistive superconducting fault current limiter and ANN-based detection

Abstract

Ensuring reliable fault diagnosis and rapid recovery in distributed generator (DG)-connected distribution systems is critical, as the integration of DG sources significantly elevates fault current levels. This study proposes an integrated approach that combines a resistive superconducting fault current limiter (RSFCL) with an artificial neural network (ANN)-based intelligent fault diagnosis framework. The objective is to limit excessive fault currents while improving detection accuracy under varying network configurations. The RSFCL is strategically placed by analyzing fault current magnitude, voltage quality, and resistance value to achieve effective current limitation without compromising system stability. Meanwhile, the ANN employs symmetrical components of current and voltage as diagnostic features. To enhance robustness, correlated variables are identified and eliminated during feature selection, strengthening the model’s fault discrimination capability. Simulation results demonstrate that the optimal RSFCL placement reduces fault current contribution ratios by up to 82.16% under symmetrical fault conditions. The ANN-based fault detection model achieves a validation accuracy of 99.7%, outperforming conventional threshold-based methods by minimizing nuisance tripping and improving circuit breaker coordination. Overall, the combined RSFCL–ANN framework provides an effective and intelligent solution for fault diagnosis and protection in DG-integrated power systems.