Software development of optimal substation ground grid design based on genetic algorithm and pattern search

Abstract

Substation ground system serves to safety of personnel and as major equipment during earth faults, which deserves considerable attentions. Basic substation safety assessment quantities include ground grid resistance, mesh touch potential and step potential, moreover, optimal design of substation ground system should consider both safeness and cost. To fill the lack of such suitable and economical optimal design software in North American industry, a software package coded in MATLAB is developed and its core algorithm and main features are introduced in this paper. A novel hybrid GA-PS optimization or two-steps optimization method is developed, in which Genetic Algorithm (GA) is used firstly to search an approximate start point range for the further optimal searching, followed by Pattern Search (PS) to find the final optimal result. This software can give analysis of ground grid safety performance and is able to recommend optimal grid design for given safety constraints. In order to make sure the accuracy of this software, the results of grid safety assessment obtained here are also compared with the results calculated by using worldwide used software WinIGS.