Power grid support from battery storage paired with solar panels under lower compensation for electricity exports

Distributed photovoltaic (PV)-battery portfolios can reshape bulk-system operation under net billing, yet their schedules reflect prosumer rather than system objectives. This study develops an export-enabled bilevel production cost model (PCM) coupling transmission-constrained unit commitment with aggregated prosumer virtual power plants (P-VPPs). Each P-VPP solves a continuous linear program under asymmetric import and export remuneration. Strict import–export coefficient ordering makes optimal grid exchange direction-exclusive, allowing exact lower-level embedding through primal feasibility, dual feasibility, and strong duality. This yields a mixed-integer linear program (MILP). A complete-year rolling-horizon study on an Australian National Electricity Market-inspired system compares matched sequential and integrated formulations across battery and remuneration scenarios. Across battery-enabled cases, integrated selection among P-VPP-optimal schedules reduces annual production cost by 2.74%–7.08% relative to the sequential benchmark, peak scheduled net demand by 0.81–2.02 GW, and mean absolute hourly variation by 10.9%–62.0%. It also raises utility-scale variable renewable energy utilisation by 1.50–3.42 percentage points and reduces synchronous-unit starts by approximately 13%–50%. The best-performing P-VPP case reduces annual production cost by 23.6% relative to the conventional PCM reference without the P-VPP portfolio; this combined effect includes distributed PV, storage, remuneration, and integrated selection. Under the optimistic selection convention, the integrated formulation provides an optimistic rolling-horizon benchmark while preserving P-VPP optimality. Larger batteries increase absolute storage use but reduce the fraction of centralised incremental production-cost value realised, showing that system value depends jointly on capacity, remuneration, and schedule-selection architecture.