A novel zero voltage transition boost converter and artificial neural network‐based estimation of converter efficiency
Naim Süleyman Tınğ, Fulya Aslay, Yakup Şahin
Abstract
Abstract In this paper, a novel zero voltage transition (ZVT) boost converter is proposed, and the overall efficiency of the converter is predicted with an artificial neural network (ANN) model. In the proposed converter, the main switch is turned on by ZVT and turned off by zero voltage switching (ZVS). Also, the other semiconductor elements operate by soft switching (SS). Besides, the proposed snubber cell has the bidirectional direct power transfer feature. The theoretical analyzes of the converter are verified by an prototype having 50 V DC input voltage, 100 V DC output voltage, 250 W output power, and 100 kHz switching frequency. The overall efficiency of the converter in hard switching (HS) condition is increased from 87.2% to 95.4% thanks to proposed snubber cell. Moreover, the efficiency of converter at HS operation is estimated with ANN. For this estimation, 110 efficiency values are obtained based on the different switching frequency and the output power values. When the actual efficiency measurements and the estimation results obtained with the ANN model are compared, it is seen that the results overlap and is obtained very close result to the truth by ANN. Thus, owing to the ANN model, the semiconductor power elements will not need to be operated at high frequencies and overheating, and the damaging to the elements will be prevented. Finally, the efficiency curve measurement of the converter takes long time in the experimental study when it takes highly short time as a few minutes in the estimation with ANN.