Harmonic current estimation using blind signal processing techniques

Blind Source Separation methods are used for feature extraction in biomedical and image processing applications. An increased use of non-linear loads results in generation of harmonics, which cannot be easily identified in electric power systems. Using blind source separation methods like fastICA and efficient fastICA the harmonic currents are estimated in an interconnected system. In this work, a performance evaluation between these two methods for harmonic state estimation is done in the form of recording and analyzing the miniscule error that exists between the actual and estimated harmonic currents. The graphical results of a simple five bus system for the two methods are also discussed.