Feature extraction for load identification by means of CPC in smart grid

In this a method for new feature extraction of non-linear loads is presented. The method is applicable to identifying harmonic loads. Namely, non-linear loads implemented in power systems can be analyzed either as a periodical disturbance or transient ones. The paper focuses on the harmonic spectral decomposition of periodic waveforms. The theory is described as an extension to the Currents' Physical Components Theory, which is a method to define various physical components in a given current waveform. In this paper, all nonlinear loads are assigned to have a unique signature and are considered as a power quality problem. Then the theory is implemented on simulated waveforms, as well as measured ones attained from power quality monitors.