高效的光伏发电预测以实现中国碳中和

Abstract Rational utilization of photovoltaic (PV) power generation is a key pathway for China to achieve carbon reduction. However, many physics-based prediction methods using climate data have not fully accounted for the significant discrepancies and data biases in Global Climate Models (GCMs) across different regions. To address this issue, a novel PV power prediction framework based on near-surface air temperature and solar radiation is proposed. A region-divided and period-segmented improved Delta method is also proposed to significantly reduce the simulation errors of climate data by coupling with Bayesian Model Averaging (BMA). The conversion efficiency of PV cell and four Sharing Socioeconomic Pathways are considered collaboratively for efficient PV power potential prediction. Nearly six million climate data volumes were quantitatively analyzed, demonstrating the strong applicability of this approach. The relevance coefficients for surface downwelling shortwave radiation and near-surface air temperature increased by 11.44 % and 2.07 %, while the root mean square errors decreased by 53.86 % and 56.21 %. The prediction results show that China is more favorable to PV power generation under the sustainable development path during 2030–2059 and 2060–2099, the average value of PV power potential can reach to 110.86 W m −2 and 168.51 W m −2 , leading to carbon emission reductions of 0.968 t m −2 yr −1 and 1.472 t m −2 yr −1 . If China’s installed PV capacity reaches 15,000 km2 and PV cell conversion efficiency rises to 50 %, carbon neutrality could be realized during 2044–2059. This study provides a theoretical basis for efficient PV power prediction and energy policy formulation in China, while also offering a methodological support for other countries with similar demand.