The number of power electronics devices connected to the power grid has increased significantly during the last years. The three main reasons for a constantly increasing number of applications for power electronics are the required increase in efficiency, renewable power sources and reduction of semiconductor prices. New wide bandgap transistors have been introduced to increase switching frequency of power electronics devices by a factor of ten in the last decade, and switching frequencies for some power electronics converters have already reached 1 MHz. This introduces several potential EMI issues. Firstly, the constantly rising number of power electronics devices and growing switching frequencies can affect total grid robustness and other devices nearby. Secondly, the normal operation of equipment connected close to the interference source can be disturbed. As a result, passive EMI filters have already become a mandatory component for power electronic devices.
The normal level of abstraction and simplification, which gives precise simulation results for the majority of applications, is usually not sufficient for designing and modelling EMI filters. The majority of the existing design procedures of EMI filters can be referred to as structured “trial and error” methods. They require a “looped” design and testing procedure. Developed EMI filters have to be tested together with a PE converter until the level of interference emitted by this converter complies with particular norms. Obviously this approach can lead to an unlimited number of iterations and very high development costs. Moreover, every iteration requires prototyping of the previously simulated EMI filter in order to improve the accuracy of the model. In this way, mutual couplings and nonlinear permeability effects are extracted from a built prototype and an equivalent circuit of the filter is fine-tuned in accordance with measurements.