A Timer and Mixed Integer Linear Programming Load Shedding Scheme for Resilient DC Microgrids

Abstract

Islanded DC microgrids are vulnerable to voltage instability caused by excessive power demand, which can adversely impact downstream consumers and disrupt overall microgrid operation. Existing load-shedding techniques face limitations such as over-shedding due to fixed voltage thresholds and time delays, predetermined load-shedding actions that fail to account for disturbance magnitude, and delayed stabilization caused by sequential load-shedding steps. To address these challenges, this paper proposes a novel load-shedding strategy for islanded DC microgrids that integrates a short-timer mechanism with Mixed Integer Linear Programming (MILP) optimization. The proposed approach reduces reliance on communication systems and achieves optimal load-shedding decisions using local voltage measurements. Simulation results on a DC microgrid model adapted from the IEEE 37-bus network demonstrate the effectiveness of the proposed scheme. The scheme results in a 20% reduction in unnecessary load shedding, an 18% improvement in voltage stabilization (measured as the final voltage after a disturbance), and a 25% decrease in response time compared to conventional methods. The results show that the proposed strategy ensures that the DC bus voltage remains above the critical threshold of 720 V, enhancing system reliability by minimizing voltage transients, reducing regulation time, and maintaining power balance. These improvements highlight the potential of the proposed scheme to support robust, secure, and resilient DC microgrid operation.

Keywords

• DC microgrid
• distributed energy resources
• load shedding
• mixed-integer linear programming
• optimization
• resilience