A fault detection method based on active injection using dual active bridge converter in photovoltaic DC integration system

Abstract

Fault characteristics in photovoltaic DC integration systems exhibit short durations, making traditional protection principles less effective and unreliable. To address this challenge, a control-protection coordinated fault detection strategy based on active dual-frequency signal injection is proposed. By modifying the phase shift modulation of DAB converters, two specific frequency signals are injected with flexible control over frequency, amplitude, and duration according to system requirements. The fault characteristics under injected signals are subsequently analyzed, revealing that for healthy lines, the calculated capacitance closely matches the actual values, while for faulty lines, it consistently exceeds the true value. Based on this observation, a fault line identification method utilizing the Relative Capacitance Deviation (RCD) algorithm and fluctuation criteria is developed. Furthermore, a high-precision fault location method is introduced, employing parameter identification combined with a nonlinear interior-point optimization algorithm. The proposed method requires no communication or prior knowledge of line parameters, and demonstrates robust performance under transient resistance up to 200 Ω and noise levels up to 20 dB. Simulation results validate the effectiveness of the proposed method.