Power loss estimation of electrolytic capacitor using genetic algorithm

In this paper, the authors propose firstly a method based on genetic algorithm to identify the parameters of a novel electric equivalent schema of electrolytic capacitors. Genetic algorithms allow to find the parameters of the model starting from measurements of magnitude and phase according to the frequency and different temperatures, which are obtained using a HP Impedance-Analyzer. Then, they show that the main advantage of this model is that all its parameters are frequency independent and it also redraws with high accuracy the capacitor behaviour for several temperatures and for a large frequency range. Finally, the identified model is introduced into a Boost Power Factor Correction circuit. The power loss variation versus switching frequency is computed using PSIM Software for different temperatures. By the end of this work, we propose the new power loss calculation approach which is more accurate than those already used by taking into account the increase of ESR at high frequencies due to the Skin Effect.