Prediction of Electromagnetic Compatibility problems based on artificial neural networks

Prediction of electronic equipments and systems’ Electromagnetic Compatibility (EMC) issues at early develop stages is inevitable for achieving their EMC discipline. This paper proposes a method to fast predict EMC problems based on Artificial Neural Networks (ANN). By means of choosing relevant Electromagnetic Interference parameters to compose of input prediction features, using the back propagation (BP) neural network to construct the mapping relation between the input prediction features and the electromagnetic disturbance response of the sensitive system, using training sample set that computed by the electromagnetic computational method to train the BP networks, and then the EMC fast prediction BP model is obtained. Finally, this method is used to predict the crosstalk coupling between wires. Results show that the proposed method is effective.