Design of grounding grids in transformator stations by using heuristic algorithms

In this thesis, the analysis of transmission substation grounding system and its design have been done. Genetic Algorithm (GA) and Differential Evolutionary Algorithm (DGA) methods that are intuitional algorithms are used for analysis and design. Different initial populations are used in GA method. Random, Halton, Hammersley, Faure, Sobol, Ortogonal, Regular, Logistic Chaotic Map, Uniform are used as initial population. The performances of GA in which these sequences are used for analysis and design of grounding system are determined. For grounding system uniform and two-layer soil models are examined. Thus, the seasonal analysis is also done. New models are formed for the seasonal analysis, GA and DGA are applied.In the design of grounding grid, both the safety condition and minimum cost are provided. It is the first time that intuitional algorithms have been applied to two-layer soil model in this thesis. The optimization of grid burial depth is done. Thus, it is determined which layer the grounding grid should be located in two-layer soil.Grid inductance and impulse impedance analysis are done for grounding system and grounding grid is designed via GA.In two-layer soil model, the effect of grounding rods is determined by using heuristic algorithms.For the problem of grounding grid design Matlab optimization toolbox and GA are compared. It is shown that GA is able to reach the solution without related to initial values.For grounding grid, the effects of conductor types are discussed and grounding grids are designed by using different conductors via GA.