Indian Journal of Science and Technology

Abstract

Load composition data is used in different static and dynamic studies of present power systems. Increasing the number and capacity of power electronic loads in the modern power systems, especially microgrids, makes it necessary to perform various harmonic studies, beside considering this new load type in the conventional studies. For this purpose, new methods are required to decompose power electronic loads from conventional loads of power systems. This paper introduces a novel non-intrusive decomposition method, based on digital wavelet transform. The contribution of the proposed method is its ability to decompose most common power electronic loads, including uncontrolled/semi-controlled rectifiers with both constant power and resistive DC loads, in steady state and especially in transient and asymmetric conditions. Moreover, this method is significantly easier and more accurate than available machine learning-based decomposition methods. Another advantage of this paper is employing excellent denoising ability of digital wavelet transform, compared to conventional digital filters, for noise elimination of actual measured waveforms. The validity of proposed decomposition method is evaluated by simulation of dominating rectifiers in combination with induction motor and inductive load, as conventional load, in MATLAB/Simulink. Laboratory-implementation of the test system is also employed for further validation of the proposed method.
Keywords: Denoising, Discrete Wavelet Transform, Line-Commutated Rectifiers, Load Decomposition, Microgrids, Transient Conditions