PERFORMANCE EVALUATION OF HARMONIC CURRENT ESTIMATION USING INDEPENDENT COMPONENT ANALYSIS
Article Sidebar
Published
Jul 6, 2012
Download
Statistic
Main Article Content
Abstract
An augmented use of non-linear loads results in generation of harmonics, which is difficult to estimate in electric power systems. Using, Independent Component Analysis (ICA) methods like
fastICA (FICA)the harmonic currents can be estimated without the knowledge of the topology of the system, particularly when the probability distribution of majority of the harmonic sources is non-gaussian. A typical scenario with all the harmonic sources close to Gaussian is considered and hence efficient variant fastICA (EFICA) algorithm is attempted in the paper. A harmonic state estimation using ICA is implemented on an IEEE 14 bus system and on a laboratory three bus model. A performance evaluation between the two methods of ICA for harmonic current estimation is done in the form of recording and evaluating the infinitesimal error that exists between the actual and estimated harmonic currents for the two interconnected bus systems. The graphical results of both the systems for the two methods also point to the superior performance of EFICA algorithm.
fastICA (FICA)the harmonic currents can be estimated without the knowledge of the topology of the system, particularly when the probability distribution of majority of the harmonic sources is non-gaussian. A typical scenario with all the harmonic sources close to Gaussian is considered and hence efficient variant fastICA (EFICA) algorithm is attempted in the paper. A harmonic state estimation using ICA is implemented on an IEEE 14 bus system and on a laboratory three bus model. A performance evaluation between the two methods of ICA for harmonic current estimation is done in the form of recording and evaluating the infinitesimal error that exists between the actual and estimated harmonic currents for the two interconnected bus systems. The graphical results of both the systems for the two methods also point to the superior performance of EFICA algorithm.