Parameter Identification of Complex Power Quality Disturbance Signals Based on Two-stage Search

Abstract

Translate Article

To address the challenge of quickly and accurately identifying the characteristic parameters of complex power quality disturbance signals in noisy environments, a parameter identification method for atomic decomposition based on two-stage search is proposed by partitioning the damped sinusoidal atom library according to different types of disturbances. In the coarse search, the fundamental component is separated from the disturbance signal, and then the rough values of the start time, frequency, phase, attenuation coefficient, and amplitude of the residual disturbance signal are quickly obtained. In the fine search, a pseudo-Newton method is used to improve the search strategy of the matching pursuit (MP) algorithm, optimizing the disturbance characteristic parameter values determined by the coarse search to obtain the best atoms matching the disturbance type. The atom library includes fundamental, quasi-fundamental, impulse, fluctuation, oscillation, and harmonic. To confirm the effectiveness of the presented approach, parameter identification is conducted on different types of composite disturbances in noisy environments. Simulation findings demonstrate that the presented approach ensures parameter identification accuracy while providing good real-time performance and noise resistance.