Using neural networks for predicting the integrated circuits susceptibility to conducted electromagnetic disturbances

Abstract

The decrease in normal signal levels, the increase in the operating frequencies, and the use of digital electronics in modern systems have led to the need for heightened EMC considerations, and consequently, better EMC models. A preliminary mathematical model based on neural networks theory is developed for predicting the level of susceptibility of integrated circuits to conducted electromagnetic disturbances such as a sine wave. A good correlation between measured and simulated results is obtained.