Analysis of Cancellation Path Estimation Errors in Narrow-band Adaptive Digital Active EMI Filters

Fast switching power electronic systems can be critical emitters of electromagnetic disturbances. For the reduction of the disturbances, active EMI filtering (AEF) promises reduced overall weight and volume compared to passive filters. Analog and digital AEF approaches are known. With help of fast and precise analog-to-digital converters, digital-to-analog converters and modern digital signal processing hardware as, e.g., field-programmable gate arrays (FPGAs), digital active EMI Filters (DAEFs) can realize complex filtering algorithms and reach higher frequencies than analog approaches. Until now, digital adaptive notch filters or narrow-band adaptive digital active EMI filters (NADAEFs) have been successfully applied to different applications. But for stable and successful operation, a valid estimation of the transfer function of the cancellation path is needed. In this paper, cancellation path estimation errors are investigated with regard to their impact on active EMI reduction. A transfer function model is presented and used for a detailed performance analysis in order to find critical parameters and potential for optimization.