Un-terminated Black-Box EMI Models of Power Converters Driven by Random Modulation Strategies

Abstract

Due to the development of renewable energy systems, there is an increasing need for tools to predict the Conducted Emissions (CE) exiting from power inverters, so to provide guidelines for optimal filter design. The objective is not only to limit the high-frequency noise injected into the distribution network (ac side) but also to protect the electronic circuitry on the dc side of the inverter. Hence, it is necessary to derive Electromagnetic Interference (EMI) models of the power inverter at both sides of the device. To this end, the (ac-side) behavioral model presented in previous works is here extended to predict the CE exiting both sides of a power inverter interfacing a PV panel to the distribution grid. Besides, suitable conditions for applying black-box modelling in the presence of random modulation strategies are investigated. It is proven that model effectiveness mainly depends on the base switching frequency and frequency band, while the effect of randomness does not significantly degrade prediction accuracy.