Edge cutting and AI-driven protection strategies for DC microgrids: A comprehensive survey of challenges, technologies, and future trends

Document Type

Review

Publication Date

6-1-2026

Abstract

Direct Current (DC) microgrids are emerging as a transformative solution for efficient energy distribution, especially with the growing integration of renewable energy sources and DC-native loads. However, their distinct operational characteristics, such as the absence of natural current zero-crossing, rapid fault current escalation, and bidirectional power flow, introduce significant protection challenges. Existing reviews on DC microgrid protection strategies primarily focus on conventional overcurrent and impedance-based methods, yet they often neglect AI-integrated, adaptive, and cyber-secured protection frameworks, as well as breaker-less converter coordination and solid-state transformer applications. Comparative evaluations across voltage levels and network topologies also remain limited, restricting practical implementation insights. This paper addresses these gaps by providing a comprehensive analysis of both traditional and advanced DC microgrid protection schemes, including derivative-based, AI-driven, and hybrid methods. It also explores the integration of edge intelligence, cybersecurity, and adaptive coordination, which collectively enhance fault detection speed, selectivity, and system resilience. These integrations enable faster response, secure communication, and dynamic adaptation for changing network conditions, ensuring more reliable and intelligent protection for modern DC microgrids. The findings of this study will benefit researchers, engineers, manufacturers, and policymakers seeking to develop intelligent, secure, and interoperable protection systems for next-generation DC microgrids.

Keywords

DC microgrid, Deep learning, Edge AI, Machine learning, Microgrid protection, Smart grid, Solid-state transformer

Publication Title

Results in Engineering

ISSN

2590-1230

DOI

10.1016/j.rineng.2026.110013

Volume

30

Publisher

Elsevier

Share

COinS