TY - GEN
T1 - Study of blind source separation on transmission line EMI
AU - Li, Zi Hua
AU - Xiao, Chun Yan
AU - Gao, Shuai
PY - 2014
Y1 - 2014
N2 - The transmission line is an important part of electrical system. Electromagnetic interference (EMI) signals can be picked up by transmission lines in a way of conduction or radiation, and interfere the sensitive devices located in the power source end and the load end, so it is important and necessary to separate and identify the EMI source signals coupling to the transmission lines in order to guide the electromagnetic compatibility (EMC) design and the further EMI diagnosis and suppression. Fast independent component analysis (FastICA) algorithm is studied and programmed, and its feasibility and separation performance are validated via simulation of BSS of three mixed signals and the average signals to interference ratio (SIR) is approximately 30 dB. The model of crosstalk of transmission lines is built and simulated, the interference signals are separated by the FastICA algorithm, and the average SIR is over 20 dB. Periodicity and spectral characteristics of the separated interference signals are analyzed, and the identification of interference signals is realized.
AB - The transmission line is an important part of electrical system. Electromagnetic interference (EMI) signals can be picked up by transmission lines in a way of conduction or radiation, and interfere the sensitive devices located in the power source end and the load end, so it is important and necessary to separate and identify the EMI source signals coupling to the transmission lines in order to guide the electromagnetic compatibility (EMC) design and the further EMI diagnosis and suppression. Fast independent component analysis (FastICA) algorithm is studied and programmed, and its feasibility and separation performance are validated via simulation of BSS of three mixed signals and the average signals to interference ratio (SIR) is approximately 30 dB. The model of crosstalk of transmission lines is built and simulated, the interference signals are separated by the FastICA algorithm, and the average SIR is over 20 dB. Periodicity and spectral characteristics of the separated interference signals are analyzed, and the identification of interference signals is realized.
KW - Blind source separation (BSS)
KW - EMI
KW - Fast independent component analysis (FastICA)
KW - Transmission line
UR - https://www.scopus.com/pages/publications/84891593713
U2 - 10.4028/www.scientific.net/AMR.846-847.493
DO - 10.4028/www.scientific.net/AMR.846-847.493
M3 - 会议稿件
AN - SCOPUS:84891593713
SN - 9783037859391
T3 - Advanced Materials Research
SP - 493
EP - 499
BT - Advances in Mechatronics, Automation and Applied Information Technologies
T2 - 2013 International Conference on Mechatronics and Semiconductor Materials, ICMSCM 2013
Y2 - 28 September 2013 through 29 September 2013
ER -