Feasibility analysis and application of graph processing in Gauss and newton-Raphson power flow calculation

摘要

Graph processing is a new kind of parallel computing framework with superiority in dealing with massive data with graph structure and complex relevance. In the power system, the grid has a natural graph structure, and the relevance between bus states is very strong. Moreover, the rapid development and application of information technology has also brought the explosive growth of the amount of data for the power system. This paper discusses the application of graph processing in some power flow calculation methods. Firstly, the features of graph processing and the feasibility of some existing power flow algorithms in the graph processing system are analyzed in detail. Then the triangular decomposition of matrix and the forward substitution based on graph processing are proposed. Lastly, cases of different sizes are presented to verify the correctness of theoretical analysis and the features of graph processing.