A settingless fault detection approach for MVDC network

Abstract

Abstract Medium Voltage Direct Current (MVDC) distribution networks present a promising solution for modern power distribution, yet fast and selective fault detection remains a significant challenge. Conventional protection methods often rely on predefined threshold settings and struggle to distinguish between internal and external faults—especially in the presence of high-impedance conditions or outfeed scenarios. This paper proposes a setting-less fault detection algorithm that determines fault locations by evaluating the rate of change in the difference between local and remote terminal currents. The approach eliminates threshold dependency, enhancing adaptability, sensitivity, and robustness. Advanced signal processing techniques are employed to suppress line capacitance discharge effects and minimize data exchange between terminals. To validate the scheme, extensive simulations are carried out on a multi-terminal MVDC system using PSCAD/EMTDC and MATLAB. The proposed method successfully achieves fault detection times as low as 0.25 ms, accurately detects faults with resistances up to 200 Ω, and maintains robust operation under 50 dB Gaussian noise, sudden load changes, and AC-side disturbances. The results confirm the scheme’s effectiveness and suitability for real-world MVDC protection applications.