A Data-driven Modeling Method for Power Quality Disturbance sources in Distribution Networks

With the large-scale access of distributed energy, the power quality of the distribution network has attracted more and more attention. The power quality disturbance source modeling of the distribution network is the prerequisite for analyzing the influence of the disturbance source on the power grid, identifying and locating the disturbance source, and improving the disturbance source management. In recent years, power companies have extensively carried out online power quality monitoring and accumulated massive amounts of power quality disturbance source operating data, which lays a good foundation for disturbance source modeling. This paper proposes a data-driven power quality disturbance source modeling method for distribution networks, including data-driven modeling procedures, data preprocessing methods, timing and statistical characteristic parameters acquisition methods. Compared with the traditional model method based on mechanism simulation, the data-driven model has the advantages of continuous improvement and complete characteristics. Finally, an example of a photovoltaic scenario verifies that the method has good convergence and engineering practicability.