Anomaly Data Cleaning for Distributed Photovoltaic Power Plants Based on Sliding Standard Deviation
摘要
In this paper, a anomaly data cleaning method based sliding standard deviation for distributed photovoltaic power plants is proposed. Due to lack of irradiance measurement information, the DC voltage and output power were selected for data cleaning. The DC voltages were categorized into three groups, the before starting voltage group, the low irradiance operating group and the normal operating group. The method is used for abnormal data identification through grouping of DC voltages, calculation of sliding standard deviation and threshold setting. The data of a real distributed PV plants are take as a example and the results illustrate the effectiveness of the method by cleaning the abnormal data.