Research on the Architecture of Digital Enabling Detection of Data Anomalies in Intelligent Distribution Network

With the continuous advancement of the construction of new power systems, the digitization degree of the power grid has been continuously improved, and a large number of distribution network data has been collected. Multi-source heterogeneous big data has problems such as complex structure, redundancy, poor quality, etc., which is difficult to meet the requirements of the next generation power grid for the visible and controllable distribution network. The data quality evaluation and anomaly detection methods of distribution network are studied. Firstly, the missing value filling method based on MVSD and the anomaly detection method of time series anomaly detection algorithm (S-H-ESD) are proposed. On this basis, according to the principle of conservation of capacitance, the voltage correlation analysis of low-voltage user load data is carried out, and dimension reduction is carried out to achieve multi-dimensional quantitative data quality assessment and anomaly monitoring. Through the analysis of the measured data, the feasibility and effectiveness of the proposed algorithm are proved, which can improve the data quality of distribution network and provide technical support for the construction of a new generation of power grid.