A Flexible Black-Box Model for Conducted Emission Predictions with Different Switching Frequencies

Black-box modeling technique is an efficient approach to represent the electromagnetic interference behavior of power converters, whose presence may cause malfunctioning in adjacent electronic devices. Although developing a black-box model is simpler and less demanding than extracting an explicit circuit model, model effectiveness is limited to certain operating conditions, such as a fixed modulation strategy. In this work, a flexible black-box model is proposed, which can be effectively used for prediction also in case of different modulation conditions without requiring a new estimation of model parameters if modulation parameter (such as the switching frequency) changes. Flexibility is achieved by modeling time-domain noise waveforms using an analytical curve-fitting model or an autoregressive model, whose accuracy is compared in time and frequency domain. The proposed model is experimentally verified on a boost converter operated with different switching frequencies.