Construction and application of high-order tensor network models for large-scale electric power grids

With the continuous construction and promotion of ultra-high level voltage lines, the geographical scope of China's power grid is constantly expanding. At the same time, the complexity of the topology structure of the power grid, the types and quantities of electrical equipment, and the operational status of the power grid are constantly expanding and becoming more complex. However, the current data description of the operational evolution trend of large-scale power grids still remains at the stage of separately describing from the perspective of multiple application systems. In response to this issue, the article proposes a method for constructing a high-order tensor network model for large-scale power grid analysis applications. Due to the large number of equipment components in the power grid and their extremely high frequency of change, the use of intuitive description methods cannot meet the fast and real-time expression requirements of the power grid operation situation. Using the equivalence method to replace various entities in the power grid, and using numerical methods to replace a large number of components such as lines and switches in the power grid. Based on this, the vector form of power system operation data is transformed into a three-dimensional tensor form, thereby retaining the spatial distance of power system nodes and reflecting the spatial correlation of power system operation data. At the same time, it can be used as input for typical power analysis calculation models for power system analysis