DocumentCode
3258252
Title
Comparative Evaluation of Harmonic Extraction Techniques for Three-Phase Three-Wire Active Power Filter
Author
Chudamani, R. ; Vasudevan, Krishna ; Ramalingam, C.S.
Author_Institution
Indian Inst. of Technol. Madras, Chennai
fYear
2007
fDate
27-30 Nov. 2007
Firstpage
1700
Lastpage
1706
Abstract
In this paper the performance of four harmonic extraction algorithms for active power filters are evaluated based on their steady state and dynamic performance. The techniques which are considered for comparative study are (i) filtering approach using the conventional low pass filter, (ii) instantaneous real and imaginary power (pq) theory, (iii) instantaneous active and reactive currents method (dq), and (iv) nonlinear least squares (NLS) estimation method. Comparison is based on simulation studies carried out in SABER. It is shown through simulation results that the dynamic behaviour of the harmonic extraction approach has an effect on the active power filter ratings. The simulation study reveals the superiority of the nonlinear least squares approach. This method provides good dynamic response and also extracts the load current harmonics explicitly.
Keywords
active filters; least squares approximations; power harmonic filters; harmonic extraction approach; harmonic extraction techniques; imaginary power theory; instantaneous real theory; low pass filter; nonlinear least squares estimation method; reactive currents; three-phase three-wire active power filter; Active filters; Filtering theory; Frequency; Least squares methods; Low pass filters; Power harmonic filters; Power system harmonics; Steady-state; Switched-mode power supply; Voltage; Compensation lag; Harmonic elimination; Nonlinear least squares;
fLanguage
English
Publisher
ieee
Conference_Titel
Power Electronics and Drive Systems, 2007. PEDS '07. 7th International Conference on
Conference_Location
Bangkok
Print_ISBN
978-1-4244-0645-6
Electronic_ISBN
978-1-4244-0645-6
Type
conf
DOI
10.1109/PEDS.2007.4487937
Filename
4487937
Link To Document