Optimizing energy supply-demand in photovoltaic greenhouses: A three-layer framework informed by the crop production calendar

The energy demand of crops and the on-site photovoltaic (PV) supply in greenhouses vary dynamically across different growth stages. While PV systems have been applied to greenhouses for energy saving, their design and operation are rarely coupled with the stage-specific characteristics of crop growth. To bridge this gap, a supply–demand matching coefficient was developed and applied to a year-long tomato cultivation experiment, quantifying significant mismatches. The results reveal markedly different energy alignment across stages, with particularly abrupt mismatches at fallow boundaries. This pattern challenges the conventional data-driven-based PV-battery optimization approach, which lacks mechanisms to handle such stage-wise transitions efficiently. To address this, a novel three-layer optimization framework is proposed wherein the scheduling and control schemes are actively informed by the crop production calendar, thereby systematically coordinating capacity design, weekly scheduling, and real-time control with growth stages to eliminate the adaptation inefficiencies inherent in the conventional approach. Comparative results demonstrate that, when benchmarked against a grid-only greenhouse, the proposed optimization framework achieves a net present value (NPV) of 2.14 million CNY and a self-sufficiency rate of 0.269. Against the conventional PV-battery optimization approach, it delivers a significant 18.0% improvement in NPV. This study demonstrates that explicit integration of the production calendar can significantly enhance the economic and energy performance of PV-powered greenhouses, offering a replicable pathway toward cost-effective and low-carbon protected cultivation.