Specialized Genetic Algorithm to Solve the Electrical Distribution System Expansion Planning

Abstract

A specialized genetic algorithm with a adaptation of Chu-Beasley algorithm is presented in this paper to solve the electrical distribution distribution system expansion planning (DSP) problem modeled by a mixed integer nonlinear programming problem. The specialized genetic algorithm proposed in this paper starting from a initial population where all elements have a radial topology found using a heuristic algorithm and after the selection and mutation operations must also go through a local improvement in order to make the proposed solution in a feasible solution, if necessary, with respect to operational constraints. The DSP problem presented in this paper consider the circuit construction/recondutoring for different types of conductors and the substation construct/reinforcement. To evaluate the quality of the proposed methodology were used three different test systems found in the literature, 23, 54 and 136 buses systems.