Optimal management of a virtual power plant considering renewable energy resource plus energy storage unit

The global effort to address extreme carbon emissions has garnered significant attention from both public and private sectors, notably through initiatives spearheaded by the United Nations, such as the Sustainable Development Goal 7 (SDG7).Within this context, renewables are crucial for achieving the SDG7; however, the inherent intermittency and unpredictability of RES outputs impose serious reliability challenges on power systems, often resulting in unforeseen power imbalances, outages, and abrupt shutdowns. To manage these technical challenges, this study proposes a data-driven modeling approach utilizing Battery Energy Storage Systems (BESS) and Diesel Generators (DGs) in a Distributionally Robust Joint Chance-Constrained Optimization (DR-JCCO) framework, formulated through Mixed-Integer Linear Programming (MILP) to maintain convexity. This study integrates tractable algorithms, specifically the Conditional Value-at-Risk (CVaR) and H-X algorithms, to improve system performance within Integrated Energy Modeling Systems (IEMs). The findings indicate that CVaR provides efficient BESS operation and operational cost minimization without necessitating wind power curtailment, though it is associated with higher computational burden. In contrast, the H-X algorithm demonstrates superiority in terms of tractability, achieves faster BESS cycling, and results in reasonable profit output margins. Furthermore, the H-X algorithm establishes a new benchmark for enhancing existing optimization models such as REopt applied in IEM applications. Finally, Binomial Expansion (BE) has been applied to selectively reformulate a non-linear upper-level objective function, to demonstrate the robustness and efficiency of linear models solved by the off-the-shelf solvers such as Gurobi.