Power System State Estimation Guided by Neural Network Considering Power Flow Constraints

At present, China's power system has entered a new stage of development, and the proportion of new energy loads connected to the distribution network is increasing year by year. For example, with the integration of a large number of electric vehicles charging stations, not only is the network topology more complex, but the uncertainty and randomness of residents' charging time and spatial distribution will also increase the difficulty of power grid control. Neural networks not only have the advantages of high efficiency and accuracy in fitting and regressing data, but also can capture the network topology structure of information and form internal connections. However, data estimation using neural networks is purely mathematical, and the black box output is difficult to explain its operating principle. The output results may not conform to the laws of physics and electricity to a certain extent, causing unnecessary losses. In this context, this article will construct a neural network embedded with a physical power flow model based on power flow calculation to estimate the state of the power system grid.