Data Quality Improvement Method of Distributed PV Generation Based on Time Correlation and Spatial Correlation

In order to solve the problem of data anomaly and data missing in distributed photovoltaic (PV) active power, this paper proposes a data quality improvement method based on time correlation and spatial correlation. In terms of anomaly detection, back propagation (BP) neural network is used to establish a correlation model based on the correlation among active power and meteorological data on the time scale, and the difference method is used to detect and eliminate anomaly data. In terms of missing data repair, Random Forest is used to establish a correlation model based on the spatial correlation among active power of power station to be repaired and its surrounding power station's, and the direct repair method is adopted to repair the missing data. The experimental results show that the anomaly detection method based on time correlation can achieve an AUC value of more than 0.93, and the repair method based on spatial correlation can achieve more accurate repair results.