With the development of artificial intelligence

models represented by deep neural networks (DNN) have been widely applied in power grid analysis and computation. However

these models still struggle to directly process data with complex topological structures

such as power grid mapping. Graph Neural Networks (GNN) can effectively capture the relationships between nodes and the global information of graphs

addressing the limitations of traditional DNN in processing graph-structured data. This paper systematically summarizes the current applications and limitations of GNN in various power grid analysis and computation scenarios

including data generation

renewable energy and load forecasting

power flow calculation

state estimation

transient analysis

and optimal power system dispatch. This paper also analyzes the advantages and challenges faced by GNN in this domain. Finally

this paper presents an outlook on future research directions for GNN in power grid analysis and computation.