Study on Blind Source Separation of Dual/Single-channel Conducted EMI Noise
ZHAO Bo1,2,ZHAO Min1,MA Yu-ming2,SHEN Fei3
1.College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu 211106, China
2.Jiangsu Institute of Metrology, Nanjing, Jiangsu 210023, China
3.College of Instrument Science and Engineering, Southeast University, Nanjing, Jiangsu 210096, China
Abstract:Aiming to the complexity and high-cost of hardware-based separation network in conducted electromagnetic interference noise, as well as the poor anti-noise ability and measuring synchronism of traditional software-based separation approach, an EMI difference mode and common mode noise extracted strategy based on blind source separation theory is presented. Firstly, the fast independent component analysis algorithm is used for live and null line noise extraction, with a dual-channel BSS approach. Experiments demonstrate a wider signal-to-noise ratio range than RF current approaches, meaning stronger applicability. Then, the single-channel BSS approach extends extra virtual channel using continuous wavelet transform algorithm. Tests prove the separation performance depends on the choice of wavelet function, rather than the number of extended channels. Single-channel BSS approach saves economic cost and possesses wide prospect, with a little less separation ability.
[1]高攸纲. 电磁兼容总论[M]. 北京:北京邮电大学出版社, 2011: 47-59.
[2]刘鹏程. 电磁兼容原理及技术[M]. 北京:高等教育出版社, 1993: 23-35.
[3]Paul C R, Hardin K B. Diagnosis and reduction of conducted noise emissions[J]. IEEE Transactions on Electromagnetic Compatibility, 1988, 30(4):19-23.
[4]See K Y. Network for conducted EMI diagnosis[J]. Electronics Letters, 1999, 35(17):1446-1447.
[5]Lo Y K, Chiu H J, Song T H. A software-based CM and DM measurement system for the conducted EMI[J]. IEEE Transactions on Industrial Electronics, 2000, 1(4):977-978.
[6]杨小牛, 付卫红. 盲源分离理论、应用与展望[J]. 通信对抗, 2006,(3): 3-10.
[7]梅铁民. 盲源信号分离时域与频域算法研究[D]. 大连:大连理工大学, 2006.
[8]金红艳. 求解大规模优化问题的有限记忆拟牛顿法[D]. 长沙:湖南大学, 2013.