Particle Swarm Optimization vs Genetic Algorithm, Application and Comparison to Determine the Moisture Diffusion Coefficients of Pressboard Transformer Insulation

Moisture mobility inside a transformer's solid insulation can be modelled by using a diffusion model based on Fick's second law. The precision of these models is related to the so-called moisture diffusion coefficient. The experimental determination of the moisture diffusion coefficient can be a difficult task. For this reason, previous studies aimed to find a more simple experimental methodology to determine the moisture diffusion coefficients of solid cellulosic insulations. This methodology uses experimental drying curves and an optimization process with genetic algorithms (GAs) working with a drying diffusion model which is solved by the finite element method. In this article, a basic particle swarm optimization (PSO) method as an alternative to the previous optimization process by GAs was implemented and evaluated. The PSO method reduces the time spent in the determination of the moisture diffusion coefficient. Additionally, optimization by particle swarm simplified the methodology to determine the moisture diffusion coefficient because a subsequent statistical analysis, as required when GAs are used, is not necessary.