Open AccessPrediction of the High Frequency Behavior in Degraded Coaxial Connector Based on Neural Network
Author(s) -
Q. Li,
AUTHOR_ID,
Waon Ho Yi,
Jian Gao
Publication year - 2022
Publication title -
international journal of electrical and electronic engineering and telecommunications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.171
H-Index - 6
ISSN - 2319-2518
DOI - 10.18178/ijeetc.11.2.156-161
Subject(s) - artificial neural network , reliability (semiconductor) , coaxial , degradation (telecommunications) , scattering , cable gland , computer science , acoustics , biological system , artificial intelligence , physics , telecommunications , optics , power (physics) , quantum mechanics , biology
Accurate prediction of high frequency behavior for the degraded contact surface is of great significance for the reliability evaluation of the connector. A prediction algorithm of neural network is proposed to forecast the high frequency scattering parameters under different degrada-tion levels. The degraded high frequency parameters are extracted according to the developed equivalent model. Simulations are conducted to predict the scattering para-meters at the specific frequencies using the BP (back propagation) and Elman neural networks, and the prediction accuracy is further compared. Moreover, the scattering parameters at 3.1GHz to 3.5GHz are predicted for the two degradation levels, which provides the variations under higher frequency.