Harmonic Components Identification through the Adaline with Fuzzy Learning Parameter

Identification of different harmonic components of current/voltage signals is required in many power system applications e.g. power quality monitoring, active power filtering, and digital system protection. In this paper, a method based on the adaptive linear combiner (Adaline) is presented for harmonic components identification. The convergence speed and the estimation error of the Adaline are governed by the learning parameter (LP) in the weight adaptation rule of this artificial neural network. Thus, instead of a constant LP utilized in the conventional Adaline, this paper proposes the implementation of a fuzzy inference system (FIS) for suitable adjustment of the LP. Two simulation studies are conducted on the MATLAB and PSCAD/EMTDC to show the validity and performance of the proposed method.