This invention relates to the field of power supply or distribution technology, and discloses a method and
system for distributed photovoltaic power prediction based on
electricity consumption information collection. The method includes: acquiring spatiotemporal data of
electricity consumption from smart meters at multiple photovoltaic access points; quantifying
wave propagation characteristics based on cloud-induced wave spatiotemporal data, and outputting
wave propagation characteristic data; reconstructing global cloud motion using
divergence-curl regularized vector spline interpolation, and outputting global cloud
motion reconstruction data; performing cloud-induced wave
advection prediction on preset target smart meters using cloud-induced wave spatiotemporal data, and outputting cloud-induced power prediction data; and then performing ensemble probability prediction to output the power probability prediction result within the future time window. Addressing the problems of numerical
weather forecasting's difficulty in analyzing local cloud shadow movement and unstable short-term power
mutation prediction, this invention seamlessly connects wave
decomposition, propagation quantization, flow field reconstruction,
advection prediction, and ensemble probability prediction, improving the spatiotemporal consistency of short-term power prediction.