A Practical Power Prediction and Error Correction Method for Distributed PV based on the Spatial Correlation

The rapid installation of distributed photovoltaic (PV) in the distribution network has caused challenges to the load forecasting and the safety of the grid, which is urgent to realize accurate prediction of distributed PV at a small cost. A practical power prediction method for low-voltage distributed PV based on the spatial correlation was proposed. Firstly, a reference PV station (RPS) was selected based on the spatial correlation principle of adjacent regions, and the electricity output correlation between distributed PV and RPS was established through historical output data. Then, the predicted power of distributed PV was calculated with the power prediction data of the RPS considering the output correlation. Thirdly, three kinds of error sources related to the proposed method were analyzed. Based on the error analysis, an error correction method for capacity constraint and generating performance difference between distributed PV and RPS was proposed. Some typical distributed PVs in Chengde were demonstrated to verify the effectiveness of the method.