Second by Second Prediction of Solar Power Generation Based on Cloud Shadow Behavior Estimation near a Power Station

Photovoltaic (PV) power generation has a particular problem for grid cooperation in that output can fluctuate due to the shadows created by clouds. If we can grasp the behavior of cloud shadows beforehand, then it may be possible to forecast output fluctuations. In this study, we want to prove if it is possible to calculate power output variation from the accumulated cloud shadow data. Cloud shadow behavior was measured from the ground by photodiodes (PDs) and the cloud shadow vector was calculated from the position and time difference. The time from the calculated cloud shadow vector to the arrival of the cloud shadow and the power generation output was calculated and compared with the actual solar power generation output. Thus, we confirmed that we can predict power generation output from a high correlation of two outputs. We found that prediction is possible, with high precision, at a short distance.