Dimensional Reduction on an Intelligent Model for Efficiency Improvement of Switching Modes Detection

This research implements a dimensional reduction with the aim of improving the efficiency of a classification algorithm used for detection of different operation modes of a buck converter. The analysis of a half-bridge buck converter is done showing two different working state: hard-switching and soft-switching. A model for dimensional reduction is used on the input data of a classification model. The dimensional reduction helps to reduce the computational costs and improve the performance of the classification model. Very good results were obtained and an improve in the classification accuracy is achieved.