PWM Signal Prediction for Broadband Active CM EMI Suppression of a GaN Motor Inverter with Adaptive FIR Filters

To increase power density and reduce system weight, active EMI filters (AEF) are investigated for power electronics in aerospace applications. For CM EMI, methods injecting cancellation pulses based on the PWM signals can be used. However, synchronous application of the cancellation pulse with the disturbance pulse can be challenging, especially for existing setups where the PWM signals cannot be intercepted and delayed before reaching the power electronics. Therefore, in this contribution, a PWM prediction method is presented, which uses an FPGA to reconstruct the PWM signals of a motor inverter. By this, the implementation of delay-compensated digital control signal DAEF (DDC-DAEF) is realized without delaying the original PWM signals. The prediction performance is investigated for multiple inverter frequencies, showing a high accuracy. Additionally, noise suppression performance is evaluated with DDC-DAEF for a motor inverter setup, comparing the results to the previous method of delaying the PWM signals.