Image Description

List of Publications by Department for the Academic Year A robust reference extraction method for shunt active power filters in the presence of noise and source voltage distortion

Abstract

Active filters are commonly employed to mitigate voltage and current distortions by rapidly responding to harmonic disturbances originating from the load side. However, the effectiveness of these filters depends on the accuracy of reference signal extraction, which is susceptible to measurement noise and source voltage distortion. This paper proposes an extraction method consisting of decomposing the distorted signal into its constituent harmonic components by the matrix pencil method, reconstructing the fundamental component from the estimated fundamental frequency and amplitude, and subtracting the fundamental component from the distorted signal to obtain the reference signal. The robustness of the proposed method to noise and source voltage distortion stems from the over-modeling technique of the matrix pencil method and its ability to extract the reference signal without a low-pass filter, which often introduces errors in phase shift and magnitude. Simulation scenarios developed in MATLAB/Simulink involving measurement noise, source disturbance, and asymmetrical loads are used to compare the performance of the proposed method with the instantaneous reactive power theory and synchronous reference frame methods. Additionally, an experimental prototype is developed to validate the effectiveness of the proposed method on real data. Analysis of simulation and experimental results shows that the proposed method outperforms the conventional methods in robustness to source distortions and measurement noise, as demonstrated by the total harmonic distortion, which is around 0.77 % for simulated data and 1.68 % for experimental data. These findings underscore the significant enhancement in the overall power filter performance facilitated by the proposed extraction method.

Journal/Conference Information

Measurement,DOI: 10.1016/j.measurement.2024.116243, ISSN: 02632241, Volume: 242, Issue: Part E, Pages Range: 0-17,