This paper proposes a monitoring technique for DC-DC converters based on time domain measurements of voltages and currents of passive components. The main goal of this work is to identify critical situations before they lead to a total loss of functionality. The subjects of this study are both the passive and active components. The analysis method can be easily applied on different topologies while maintaining the same theoretical basis. In this case a Cuk converter is used to demonstrate the effectiveness of the method and the performances of different classifiers are presented. The proposed comparison shows that the best classification results are obtained through a Support Vector Machine and a Multilayer neural network with Multi-Valued Neurons.
Smart monitoring of DC-DC converters / Bindi M.; Talluri G.; Lozito G.M.; Luchetta A.; Piccirilli M.C.; Grasso F.. - ELETTRONICO. - (2022), pp. 1-6. (Intervento presentato al convegno 2022 IEEE International Conference on Environment and Electrical Engineering and 2022 IEEE Industrial and Commercial Power Systems Europe (EEEIC / I&CPS Europe)) [10.1109/EEEIC/ICPSEurope54979.2022.9854667].
Smart monitoring of DC-DC converters
Bindi M.;Talluri G.;Lozito G. M.;Luchetta A.;Piccirilli M. C.;Grasso F.
2022
Abstract
This paper proposes a monitoring technique for DC-DC converters based on time domain measurements of voltages and currents of passive components. The main goal of this work is to identify critical situations before they lead to a total loss of functionality. The subjects of this study are both the passive and active components. The analysis method can be easily applied on different topologies while maintaining the same theoretical basis. In this case a Cuk converter is used to demonstrate the effectiveness of the method and the performances of different classifiers are presented. The proposed comparison shows that the best classification results are obtained through a Support Vector Machine and a Multilayer neural network with Multi-Valued Neurons.I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.