Intelligent Technique Proposed for Nonlinear Inductor Modelling for DC/DC Converters

The inductor is one of the main components for converting energy within power converters. The inductance value does not always be constant as it decreases with the current, particularly in the saturation region where the current’s value increases. Thus, the inductor decrease affects the dynamic of the power converter. To exploit its nonlinear behavior and to achieve a precise dynamic behavior of the converter, an accurate inductor model must, therefore, be considered. Several models has been dedicated to analyze these characteristics, but they are complicated and difficult to be implemented. In this paper, the Adaptive Neuro-Fuzzy Inference system (ANFIS) model is used to solve the nonlinear behavior of the inductor. The proposed ANFIS model is a clever neuro-fuzzy technique, which has proved to model and control diverse processes that are nonlinear. The main feature of this model is that the evolution of the flux vs current curve behavioral analysis has been taken into account using an experimentally characterized set of data. The study is set up on the DC/DC buck converter, using Matlab/Simulink. The results obtained by simulation provide the inductor current and voltage prediction, which have good agreement with the experimental measurements with relative error about 25%.