Fault Detection Algorithms in Medium-Voltage Direct-Current (MVDC) Grids

Developing reliable protection systems is critical for the advancement of medium-voltage direct-current (MVDC) grids. This paper highlights the significance of fault detection in MVDC grids, especially in ensuring the reliability and efficiency of renewable energy systems. This paper provides a comparative analysis of fault detection algorithms, including overcurrent, undervoltage, rate of change of current (ROCOC), rate of change of voltage (ROCOV), differential, and inductor voltage derivative methods. The performance of these algorithms is quantified by metrics such as the detection speed, accuracy, and sensitivity under diverse scenarios. The authors assess these algorithms within a multi-terminal MVDC grid designed for renewable hydrogen production, evaluating the detection speed across various fault types (bus and link faults) and conditions, including the variation in the fault location and resistance. The results reveal that the UV, ROCOC, ROCOV and LIVRD methods achieve detection speeds as high as 0.01 ms, outperforming other techniques under low-resistance fault conditions. By using uniform fault scenarios, we identify the most effective algorithms for rapid fault detection, aiming to enhance the protection strategies for MVDC grids. These findings underline critical performance differences between methods, guiding the design of tailored protection schemes that address specific fault challenges in renewable-powered grids. Additionally, the practical implications of these findings for designing resilient protection schemes in renewable-powered grids are discussed. Simulations are conducted using PSCAD/EMTDC V4.6 software, ensuring the consistency and accuracy of the performance comparison. The insights gained provide a concrete understanding of each algorithm's trade-offs, enabling informed decisions when selecting optimal fault detection methods to ensure MVDC grid reliability.