The Practical Research And Design Of Matrix Network Topology Analysis

Network topology analysis is the crucial component of Distribution Management System and Energy Management System, and it is also the basic software for Electric Power Analysis. The study about it has important practical value. As the development of modern power system, the demand for the speed of software analysis is creasing, so that it urgently requires a fast network topology analysis method.The paper takes adjacency matrix as a basis and makes deep study and analysis about adjacency matrix algorithm, and on the basis it proposes a new topology analysis method, namely quasi-square of connectivity matrix. The essence of Topology Analysis is to make sure that every node in the network is connected but it does not need to make clear how many stages are connected. Therefore, each time when a new connectivity matrix element is calculated, it can immediately update anoher matrix element and its symmetry element, and reflect the relationship that the two nodes indirectly use several nodes in the connectivity matrix as soon as possible. So it not only increases calculation speed, but also eliminates the space to save the new connectivity matrix. Besides, it only needs twice the square matrix operations to achieve full connectivity matrix all, which to a large extent promotes the analysis speed of topology.In order to improve calculation and increase topology speed, as well as meet the real-time requirement, on the basis of quasi-square of the connectivity matrix, the paper proposes quasi-square of the adjacency matrix which is faster than the former. It just needs calculating a matrix square to get connectivity matrix so as to reflect topology relationship. Then it could get the outcomes of topology analysis by reverse line scanning. Consequently, this method improves the speed of topology analysis.The two algorithm proposed in the paper both belong to the category of Matrix, and have numbers of advantage, such as strong structure, simple data organization, clear analysis process, excellent adjustability and so on. Compared with the traditional Matrix analysis method, the two algorithm avoid the numerous calculation produced by the Matrix self-multiplication. The two kinds of calculation both adopt nodes number optimization. It is numbered in sequence according to qualification degree of node, because the smaller the serial number of the node is, the more beneficial to determine the connectivity relations. Meanwhile, this two kinds of algorithms given in this paper were demonstrated from the view point of mathematics, confirming the feasibility of algorithm in theory.Large scales of validation and analysis examples indicate that, the two methods in the paper could promote the analysis speed greatly, and meet the speed requirement for real-time topology analysis.