Optimizing energy storage in microgrids under uncertainty: Balancing reliability, cost, and battery wear

Smart microgrids integrating renewable energy sources are important to sustainable energy development, but their stable and economic operation is challenged by the uncertainty and intermittency of renewable generation. This paper formulates a chance-constrained stochastic energy-management model that explicitly accounts for correlated renewable uncertainties. The model minimizes total operational cost, including grid-energy purchases and battery degradation, while ensuring high-probability satisfaction of power-balance and system constraints. A partial sample average approximation framework is developed to reduce computational complexity relative to conventional sample average approximation by avoiding excessive scenario-dependent binary variables. Numerical simulations on a modified IEEE 9-bus microgrid show that, for the same 1000-scenario case, the proposed method reduces solution time from 300 s to 45 s, corresponding to an approximately 85% reduction. Its operating cost remains within 1.5% of that obtained from a much larger 10,000-scenario sample average approximation used only as a near-exact reference solution. Battery participation further reduces operating cost by up to 15%, decreases equipment switching, and improves voltage profiles by 5%–10%. These results demonstrate an efficient approach for storage-aware microgrid scheduling under correlated renewable uncertainty, while validation on larger network topologies and longer planning horizons remains an important direction for future work.