Work

Hybrid Data-driven Distributionally Robust Optimization Framework for Optimal Pv–bess Microgrid Sizing Under Uncertainty

Year: 2026
Type: article
Abstract: Determining the appropriate capacity of photovoltaic (PV) systems and battery energy storage systems (BESS) remains a critical challenge in microgrid planning, particularly in regions where historical data are limited and renewable generation is highly uncertain. Conventional sizing approaches often depend on deterministic assumptions or predefined probability distributions, which may not accurately represent real operating conditions. This work presents a hybrid framework that combines data-driven learning with Distributionally Robust Optimization (DRO) to support reliable and cost-effective sizing of PV–BESS microgrids under uncertainty. The proposed approach utilizes available historical information together with uncertainty-aware optimization to identify system configurations capable of maintaining performance across a range of possible demand and generation scenarios. By incorporating robustness directly into the decision-making process, the framework reduces sensitivity to data limitations and unpredictable renewable energy behavior. Comparative evaluation using publicly available datasets indicates improvements in reliability, lower energy loss, and reduced lifecycle cost when compared with traditional deterministic and conventional robust optimization strategies. The findings suggest that integrating data-driven forecasting with DRO offers a practical pathway for designing resilient microgrids in rural, remote, and data-scarce environments where reliable energy access remains a significant concern.
Language: English
Open Access status: bronze
OpenAlex ID: W7225910249
FWCI:
Cites: 12
Cited by:
Related to: 10