Research on the Partition Method Based on the Observation Points of Voltage Distortion in Harmonic Optimization

With the use of a large number of power electronic equipment, the harmonics in the power grid are becoming more and more serious. There are many methods for harmonic mitigation and optimization such as Global network Optimization. When there are many nodes, the global network optimization method is more complicated. Considering the problem of large numbers of harmonic injection nodes in the power grid, this paper proposes a method of clustering the voltage distortion nodes according to correlation, and using the characteristic nodes to describe the harmonic voltage distortion in the region to mitigate harmonic voltage. This method is based on the observation data of the voltage distortion of the nodes for dynamic selection and grouping. The results show that this method meets the needs of Global network Optimization. The data processing methods used in this paper such as feature extraction methods and time series interpolation optimization can effectively reduce the data dimension, and get better optimization effect.