Work

Robust distributed generation and demand-balanced hierarchical multi-layer control strategy for real-time adaptive power management in interconnected microgrids using RTDS

Year: 2026

Type: article

Abstract: Interconnected microgrids (IMGs) improve resilience and renewable utilization, but coordinated inter-MG power exchange, storage-constrained dispatch, and reliable multi-mode operation remain challenging. This paper proposes a new real-time Hierarchical Multi-Layer Control Strategy (HMLCS) integrating primary, secondary, and tertiary control with explicit inter-MG power sharing and constraint-aware battery dispatch. Algebraic active-power balances coordinate grid-connected, interconnected, islanded, and self-healing modes according to generation–load imbalance and neighboring support capability. Battery operation is constrained by a 20–90% state-of-charge (SoC) range, 0.6 C charge/discharge limit, 70% depth-of-discharge constraint, and SoC-dependent derating. Small-signal converter models verify closed-loop stability across the four modes, including platform latency. Implementation uses a Real-Time Digital Simulator (RTDS) with 50 µs network and 2.5 µs converter time steps. Validation covers mode transitions, unbalanced and nonlinear loading, irradiance transients, prolonged islanding, and severe fault-plus-( N − 1) outage conditions. Over the 20 s sequence, PV supplies 72.95% of load energy, with battery storage, diesel generation, and grid support contributing 14.73%, 5.14%, and 7.19%. The maximum frequency excursion is 0.260 Hz, with primary arrest in 0.25 s and full restoration in 1.71 s. Current THD is 1.604% under nominal conditions and 2.492% with 30% nonlinear loading. During 90 min of islanding, 100% of critical-load energy is served with a final SoC of 21.66%. Under the severe-disturbance sequence, self-healing provides 12.4–17.6 kW of inter-MG support without load disconnection and reduces frequency restoration time by 39.0%. Energy not served is zero and the load-energy service ratio reaches 98.37%, confirming the dynamic, power-quality, storage-aware, and resilient real-time performance of the proposed HMLCS.

Source: Scientific Reports

DOI: 10.1038/s41598-026-74288-8