System value assessment of rooftop solar-plus-storage considering regional heterogeneity

Distributed photovoltaic (PV) and energy storage systems are playing an increasingly important role in facilitating renewable energy integration and enhancing system flexibility. However, their systemic impacts-particularly under decentralized decision-making and region-specific conditions-remain insufficiently quantified. This study proposes a unified evaluation framework that combines an agent-based bi-level optimization model with a Random Forest-based attribution analysis to assess the system value of distributed PV and storage systems across three dimensions: operational efficiency, carbon reduction, and non-wire alternatives value. The framework captures the interaction between user-level investment behaviors and system-level outcomes, while also identifying the key drivers behind regional disparities. Case studies reveal significant heterogeneity in system value, with minimum and average load levels, along with locational marginal price (LMP) volatility, emerging as dominant factors. Pricing mechanisms and storage costs show nonlinear effects: increasing the PV recovery price coefficient from 0.7 to 0.9 enhances system value by over 7 %, while storage cost reductions exhibit diminishing returns. These findings highlight that regional heterogeneity-particularly in rooftop potential, load profiles, and grid conditions-substantially shapes the value proposition of rooftop solar-plus-storage systems. The results offer policy-relevant insights to support more differentiated incentive schemes and deployment strategies across provinces.