A time-coupled multi-objective distributionally robust chance-constrained framework for grid resilience enhancement...

Area of Science:

• Electrical Engineering
• Operations Research
• Optimization

Background:

• Power grids face increasing threats from severe weather and attacks, necessitating resilient restoration strategies.
• Existing methods often lack comprehensive integration of logistical constraints and uncertainty quantification for mobile resources.

Purpose of the Study:

• To develop a unified framework for resilient power grid restoration using Mobile Emergency Generators (MEGs).
• To optimize the trade-off between restoration cost and grid resilience under uncertainty.

Main Methods:

• A time-coupled, multi-objective distributionally robust chance-constrained (MODRCC) framework was developed.
• The model integrates MEG logistics, islanding-feasible power flow, and Wasserstein ambiguity for uncertainty.
• A Mixed-Integer Second-Order Cone Programming (MISOCP) formulation was solved using an NSGA-II evolutionary algorithm.

Main Results:

• The proposed MODRCC framework reduced expected unserved energy (EUE) by 14-20% compared to baseline methods.
• Significant improvements in EUE were observed, e.g., a 54% decrease on the IEEE-118 case with a modest cost increase.
• The computational framework demonstrated efficient scalability for large-scale grid restoration problems.

Conclusions:

• Explicitly coupling mobility logistics with distributionally robust optimization provides operationally credible and cost-aware restoration schedules.
• The approach is suitable for enhancing grid resilience in disaster-prone regions.
• Wasserstein regularization effectively induces smooth, convex trade-offs between cost and resilience.