A general algorithm to automatically create admittance matrices for electric networks

Abstract

In this paper, a general method to automatically create admittance matrices for power electric networks is presented. This is a part of a joint project between Mississippi State University and University of Texas at Austin Center for Electromechanics (CEM). Starting with an electric network designed in the Mathworks software Simulink (SimPowerSystems) and output from Simulink as an XML file, we parse the XML file for the relevant information (blocks, connectivity, and parameters). The underlying algorithm consists of four stages: 1) Calculate the conductance values of the components in each block, 2) Build a matrix representing the connectivity of components, 3) Assign a unique number for each node, and finally 4) Calculate the admittance matrix using the data obtained from the previous steps. This is the initial work to support the development of a parallel computing tool where the power system will be solved using a tool developed by the University of Texas called CEMSolver. The second phase of this joint project will be finding the most optimal partitioned systems in such a way that the matrices of each partition are distributed to different CPUs and the total CPU usage for these subsystems will be minimal. Thus, the power system will be partitioned and each partition will have its own admittance matrix.