Stochastic optimization method for mechanical design of overhead distribution power lines

摘要

This article presents a stochastic optimization method for the mechanical dimensioning problem of overhead distribution power lines. For this purpose, an algorithm tailored for solving mixed integer programming problems based on the MI-LXPM algorithm was developed, which, associated with the unscented transform, allows the uncertainty inherent in the wind speed to be considered during the optimization process, with fewer evaluations than required by Monte Carlo methods. An on-the-fly population size adjustment algorithm was used, which allows the required number of evaluations to reach solutions to be reduced further. When compared to a traditional optimization process with deterministic parameters, the results of the proposed method present considerably superior reliability in the case studied, producing a reliable solution in 90.42% of all simulations, while the former produced only 76.39%.