As distributed energy resources proliferate across power systems, planners face mounting complexity in coordinating storage placement, sizing, and timing decisions amid uncertain futures. A new planning model tackles this challenge by treating short-term and long-term uncertainties as distinct, interconnected problems within a unified optimization framework.
The model employs chance-constrained programming to handle near-term volatility—such as solar and wind variability, demand fluctuations, and electric vehicle charging patterns—while simultaneously accounting for long-term unknowns like falling battery costs and uncertain renewable deployment rates. By integrating network reconfiguration capabilities alongside storage planning, the approach identifies opportunities to enhance system flexibility through topology changes as well as hardware additions.
Validation on an enhanced IEEE 33-bus test system and a 63-bus real-world Chinese network demonstrated measurable benefits. Compared to traditional robust planning methods, the chance-constrained approach reduced total planning costs by 5.55%. More importantly, when evaluated against deterministic (single-scenario) planning, it improved the system's ability to handle extreme cases—such as high renewable penetration or rapid EV adoption—by 19.20 percentage points.
The model determines three critical decisions for each candidate storage site: whether to build, when to construct, and what capacity to install. Network reconfiguration options provide an additional lever to alleviate congestion and reduce storage requirements, lowering overall investment needs.
Solution using commercial optimization software (Gurobi) confirmed computational tractability for realistic network sizes. This finding is significant because planners can now incorporate probabilistic constraints without sacrificing computational speed, enabling more frequent re-planning as new data arrives.
The work has practical implications for distribution utilities managing rapid renewable growth and electrification. By explicitly modeling both near-term dispatch challenges and long-term cost evolution, utilities can build resilient networks without over-investing in storage or over-constraining planning decisions to worst-case scenarios. The method aligns with evolving grid codes requiring flexibility and supports cost-effective energy transition pathways.
