Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification

Gisele Beatrice Sonfack; Philippe Ravier

doi:doi:10.11648/j.jeee.20221003.16

| Peer-Reviewed

Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification

Gisele Beatrice Sonfack, Philippe Ravier

Published in Journal of Electrical and Electronic Engineering (Volume 10, Issue 3)

Received: 21 May 2022 Accepted: 8 June 2022 Published: 20 June 2022

Views: Downloads:

Download PDF

Share This Article

Twitter
Linked In
Facebook

Abstract

The deployment of NILM systems and others embedded systems in the residential sector provides a large amount of data to better understand the electricity consumption habits of occupants in order to provide energy optimization solutions. The Fast Fixed-Point Algorithm for Independent Component Analysis (FastICA) can be used in the identification of loads through the separation of aggregated current and voltage waveforms from devices in the operating conditions that ensure the time and/or frequency independence between the sources. However, in addition to being less suitable for under-determined systems, the FastICA algorithm is sensitive to the initial weight vector, which affects the convergence of the algorithm. In order to solve this problem, a combination of Signal processing methods has been proposed to extract individual current curves representing load profiles from the single channel observation. First, the current mixed signal was decomposed using the EEMD algorithm to obtain IMFs for use in the BSS. As the number of IMFs is very large, the PCA algorithm was used to reduce the number of IMFs from n to r. Selected principal components were whitened and an over-relaxation factor was incorporated into the iterative Newton algorithm to process the randomly generated initial weight vector. The improved FastICA algorithm was used to separate the source components, selected the best current source from the mixed observation. Finally, the individual current analyzes and compares to the original signal. The advantage of this approach lies in the fact that it applies perfectly to NILM applications where very often only one observation is available, which is the aggregated signal. Moreover, it reveals the importance of the data sampling frequency for an accurate characterization of the load profile.

Published in	Journal of Electrical and Electronic Engineering (Volume 10, Issue 3)
DOI	10.11648/j.jeee.20221003.16
Page(s)	114-120
Creative Commons	This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.
Copyright	Copyright © The Author(s), 2022. Published by Science Publishing Group

Keywords

Desegregation, RobustICA, Non-intrusive Load Monitoring, Smart Grid, Bling Source Separation, FastICA, EEMD, PCA

References

[1]	Shao, H., Shi, X. H. and Li, L. Power Signal Separation in Milling Process Based on Wavelet Transform and In- dependent Component Analysis. International Journal of Machine Tools and Manufacture, 51, 701-710, 2011.
[2]	Torres ME, Colominas MA, Schlotthauer G, et al. Acom- plete ensemble empirical mode decomposition with adaptive noise. In: Proceedings of 36th IEEE inter-national conference on acoustics, speech and signal processing, ICASSP, Prague, Czech Republic, 22-27 May 2011, pp. 4144-4147.
[3]	Colominas MA, Schlotthauer G and Torres ME. Improved complete ensemble EMD: A suitable tool for biomedical signal processing. Biomed Signal Process Control; 14: 19-29; 2014.
[4]	Wei Xu, Xiang zhou Yan, Application of Single Channel Blind Separation Algorithm Based on EEMD-PCA- RobustICA in Bearing Fault Diagnosis, Int. J. Communications, Network and System Sciences, 10, 138-147, 2017.
[5]	Jason Heeris, Single Channel Blind Source Separation Using Independent Subspace Analysis, 2007.
[6]	Mengchen Zhao, et al., Single-Channel Blind Source Separation of Spatial Aliasing Signal Based on Stacked- LSTM, MPDI, 2021.
[7]	Lei YG et al. Application of the EEMD method to rotor fault diagnosis of rotating machinery. Mech Syst Signal Process 2009; 23: 1327-1338.
[8]	Wang D, et al., An enhanced empirical mode decomposition method for adaptive blind component separation of a single-channel vibration signal mixture. J Vib Control, Epub ahead of print 22 September, 2014.
[9]	Mohamed Nait Meziane. Identification d’appareils e´lectriques par analyse des courants de mise en marche. Traitement du signal et de l’image [eess.SP]. Universite´ d’Orle´ans, 2016. Base de Données COOLL.
[10]	Zhao, Bochao He, Kanghang Stankovic, Lina Stankovic, Vladimir. (2018). Improving Event-Based Non-Intrusive Load Monitoring Using Graph Signal Processing. IEEE Access. PP. 1-1. 10.1109/ACCESS.2018.2871343.
[11]	Bonfigli, Roberto Squartini, Stefano Fagiani, Marco Pi- azza, Francesco. (2015). Unsupervised Algorithms for Non-Intrusive Load Monitoring, 2015 IEEE 15th International Conference on Environment and Electrical Engineering, EEEIC 2015 - Conference Proceedings, 10.1109/EEEIC.2015.7165334.
[12]	Yi, C.; Lv, Y.; Xiao, H.; You, G.; Dang, Z. Research on the Blind Source Separation Method Based on Regenerated Phase-Shifted Sinusoid-Assisted EMD and Its Application in Diagnosing Rolling-Bearing Faults. Appl. Sci. 2017, 7, 414. https://doi.org/10.3390/app7040414
[13]	Ziyue Jia, Linfeng Yang, Zhenrong Zhang, Hui Liu, Fannie Kong, Sequence to Point Learning Based on Bidirectional Dilated Residual Network for Non-Intrusive Load Monitoring, Electrical Engineering and Systems Science, 30 May 2020.
[14]	Simon Henriet, Benoit Fuentes, Matrix Factorization for High Frequency Non-Intrusive Load Monitoring: Definitions and Algorithms, NILM workshop, November 18 2020, Online.
[15]	Ian T. Jolliffe and Jorge Cadima, Principal component analysis: a review and recent developments, Philosophical transactions of the royal society a Mathematical, Physical and Engineering Sciences, 2016.
[16]	A. Hyva¨rinen. Gaussian Moments for Noisy Independent Component Analysis. IEEE Signal Processing Letters, 6 (6): 145-147, 1999. Postscript gzipped PostScript pdf. Longer paper with proofs (Proc. ISCAS’99).
[17]	A. Hyvarinen and E. Oja (2000) Independent Component Analysis: Algorithms and Applications, Neural Networks, 13 (4-5): 411-430.
[18]	Dominic Langlois, Sylvain Chartier, and Dominique Gosselin, An Introduction to Independent Component Analysis: InfoMax and FastICA algorithms.

Cite This Article

Plain Text BibTeX RIS

APA Style

Gisele Beatrice Sonfack, Philippe Ravier. (2022). Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification. Journal of Electrical and Electronic Engineering, 10(3), 114-120. https://doi.org/10.11648/j.jeee.20221003.16

Copy | Download

ACS Style

Gisele Beatrice Sonfack; Philippe Ravier. Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification. J. Electr. Electron. Eng. 2022, 10(3), 114-120. doi: 10.11648/j.jeee.20221003.16

Copy | Download

AMA Style

Gisele Beatrice Sonfack, Philippe Ravier. Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification. J Electr Electron Eng. 2022;10(3):114-120. doi: 10.11648/j.jeee.20221003.16

Copy | Download

@article{10.11648/j.jeee.20221003.16,
  author = {Gisele Beatrice Sonfack and Philippe Ravier},
  title = {Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification},
  journal = {Journal of Electrical and Electronic Engineering},
  volume = {10},
  number = {3},
  pages = {114-120},
  doi = {10.11648/j.jeee.20221003.16},
  url = {https://doi.org/10.11648/j.jeee.20221003.16},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.jeee.20221003.16},
  abstract = {The deployment of NILM systems and others embedded systems in the residential sector provides a large amount of data to better understand the electricity consumption habits of occupants in order to provide energy optimization solutions. The Fast Fixed-Point Algorithm for Independent Component Analysis (FastICA) can be used in the identification of loads through the separation of aggregated current and voltage waveforms from devices in the operating conditions that ensure the time and/or frequency independence between the sources. However, in addition to being less suitable for under-determined systems, the FastICA algorithm is sensitive to the initial weight vector, which affects the convergence of the algorithm. In order to solve this problem, a combination of Signal processing methods has been proposed to extract individual current curves representing load profiles from the single channel observation. First, the current mixed signal was decomposed using the EEMD algorithm to obtain IMFs for use in the BSS. As the number of IMFs is very large, the PCA algorithm was used to reduce the number of IMFs from n to r. Selected principal components were whitened and an over-relaxation factor was incorporated into the iterative Newton algorithm to process the randomly generated initial weight vector. The improved FastICA algorithm was used to separate the source components, selected the best current source from the mixed observation. Finally, the individual current analyzes and compares to the original signal. The advantage of this approach lies in the fact that it applies perfectly to NILM applications where very often only one observation is available, which is the aggregated signal. Moreover, it reveals the importance of the data sampling frequency for an accurate characterization of the load profile.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification
AU  - Gisele Beatrice Sonfack
AU  - Philippe Ravier
Y1  - 2022/06/20
PY  - 2022
N1  - https://doi.org/10.11648/j.jeee.20221003.16
DO  - 10.11648/j.jeee.20221003.16
T2  - Journal of Electrical and Electronic Engineering
JF  - Journal of Electrical and Electronic Engineering
JO  - Journal of Electrical and Electronic Engineering
SP  - 114
EP  - 120
PB  - Science Publishing Group
SN  - 2329-1605
UR  - https://doi.org/10.11648/j.jeee.20221003.16
AB  - The deployment of NILM systems and others embedded systems in the residential sector provides a large amount of data to better understand the electricity consumption habits of occupants in order to provide energy optimization solutions. The Fast Fixed-Point Algorithm for Independent Component Analysis (FastICA) can be used in the identification of loads through the separation of aggregated current and voltage waveforms from devices in the operating conditions that ensure the time and/or frequency independence between the sources. However, in addition to being less suitable for under-determined systems, the FastICA algorithm is sensitive to the initial weight vector, which affects the convergence of the algorithm. In order to solve this problem, a combination of Signal processing methods has been proposed to extract individual current curves representing load profiles from the single channel observation. First, the current mixed signal was decomposed using the EEMD algorithm to obtain IMFs for use in the BSS. As the number of IMFs is very large, the PCA algorithm was used to reduce the number of IMFs from n to r. Selected principal components were whitened and an over-relaxation factor was incorporated into the iterative Newton algorithm to process the randomly generated initial weight vector. The improved FastICA algorithm was used to separate the source components, selected the best current source from the mixed observation. Finally, the individual current analyzes and compares to the original signal. The advantage of this approach lies in the fact that it applies perfectly to NILM applications where very often only one observation is available, which is the aggregated signal. Moreover, it reveals the importance of the data sampling frequency for an accurate characterization of the load profile.
VL  - 10
IS  - 3
ER  -

Copy | Download

Author Information

Gisele Beatrice Sonfack

Department of Electrical Engineering, Douala Advanced Vocational Center of Technology, University of Douala, Douala, Cameroon
Philippe Ravier

Philippe RAVIER, Orleans Polytechnics School, PRISME Laboratory, University of Orleans, Orleans, France

Download PDF

Sections

Plain Text BibTeX RIS

APA Style

Gisele Beatrice Sonfack, Philippe Ravier. (2022). Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification. Journal of Electrical and Electronic Engineering, 10(3), 114-120. https://doi.org/10.11648/j.jeee.20221003.16

Copy | Download

ACS Style

Gisele Beatrice Sonfack; Philippe Ravier. Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification. J. Electr. Electron. Eng. 2022, 10(3), 114-120. doi: 10.11648/j.jeee.20221003.16

Copy | Download

AMA Style

Gisele Beatrice Sonfack, Philippe Ravier. Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification. J Electr Electron Eng. 2022;10(3):114-120. doi: 10.11648/j.jeee.20221003.16

Copy | Download

@article{10.11648/j.jeee.20221003.16,
  author = {Gisele Beatrice Sonfack and Philippe Ravier},
  title = {Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification},
  journal = {Journal of Electrical and Electronic Engineering},
  volume = {10},
  number = {3},
  pages = {114-120},
  doi = {10.11648/j.jeee.20221003.16},
  url = {https://doi.org/10.11648/j.jeee.20221003.16},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.jeee.20221003.16},
  abstract = {The deployment of NILM systems and others embedded systems in the residential sector provides a large amount of data to better understand the electricity consumption habits of occupants in order to provide energy optimization solutions. The Fast Fixed-Point Algorithm for Independent Component Analysis (FastICA) can be used in the identification of loads through the separation of aggregated current and voltage waveforms from devices in the operating conditions that ensure the time and/or frequency independence between the sources. However, in addition to being less suitable for under-determined systems, the FastICA algorithm is sensitive to the initial weight vector, which affects the convergence of the algorithm. In order to solve this problem, a combination of Signal processing methods has been proposed to extract individual current curves representing load profiles from the single channel observation. First, the current mixed signal was decomposed using the EEMD algorithm to obtain IMFs for use in the BSS. As the number of IMFs is very large, the PCA algorithm was used to reduce the number of IMFs from n to r. Selected principal components were whitened and an over-relaxation factor was incorporated into the iterative Newton algorithm to process the randomly generated initial weight vector. The improved FastICA algorithm was used to separate the source components, selected the best current source from the mixed observation. Finally, the individual current analyzes and compares to the original signal. The advantage of this approach lies in the fact that it applies perfectly to NILM applications where very often only one observation is available, which is the aggregated signal. Moreover, it reveals the importance of the data sampling frequency for an accurate characterization of the load profile.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - Application of Single Channel Blind Source Separation Based-EEMD- PCA and Improved FastICA algorithm on Non-intrusive Appliances Load identification
AU  - Gisele Beatrice Sonfack
AU  - Philippe Ravier
Y1  - 2022/06/20
PY  - 2022
N1  - https://doi.org/10.11648/j.jeee.20221003.16
DO  - 10.11648/j.jeee.20221003.16
T2  - Journal of Electrical and Electronic Engineering
JF  - Journal of Electrical and Electronic Engineering
JO  - Journal of Electrical and Electronic Engineering
SP  - 114
EP  - 120
PB  - Science Publishing Group
SN  - 2329-1605
UR  - https://doi.org/10.11648/j.jeee.20221003.16
AB  - The deployment of NILM systems and others embedded systems in the residential sector provides a large amount of data to better understand the electricity consumption habits of occupants in order to provide energy optimization solutions. The Fast Fixed-Point Algorithm for Independent Component Analysis (FastICA) can be used in the identification of loads through the separation of aggregated current and voltage waveforms from devices in the operating conditions that ensure the time and/or frequency independence between the sources. However, in addition to being less suitable for under-determined systems, the FastICA algorithm is sensitive to the initial weight vector, which affects the convergence of the algorithm. In order to solve this problem, a combination of Signal processing methods has been proposed to extract individual current curves representing load profiles from the single channel observation. First, the current mixed signal was decomposed using the EEMD algorithm to obtain IMFs for use in the BSS. As the number of IMFs is very large, the PCA algorithm was used to reduce the number of IMFs from n to r. Selected principal components were whitened and an over-relaxation factor was incorporated into the iterative Newton algorithm to process the randomly generated initial weight vector. The improved FastICA algorithm was used to separate the source components, selected the best current source from the mixed observation. Finally, the individual current analyzes and compares to the original signal. The advantage of this approach lies in the fact that it applies perfectly to NILM applications where very often only one observation is available, which is the aggregated signal. Moreover, it reveals the importance of the data sampling frequency for an accurate characterization of the load profile.
VL  - 10
IS  - 3
ER  -

Copy | Download