Kalman-Filter-Based Multilevel Analysis to Estimate Electric Load Composition

This paper presents a novel method by the multilevel analysis (MLA) to estimate the composition rate of electric loads with respect to measured total current waveform by improving the performance of a previous estimation scheme proposed by the authors. The proposed MLA makes it possible to analyze the electric load composition rate (LCR), which indicates the portions of several typical loads connected to a single point of common coupling (PCC), with the partition of several levels. Then, the Kalman-filter (KF) algorithm is applied to solve the estimation problem of LCR based on formulation with current waveforms in each level. The effectiveness of MLA based on the KF in practice is verified by the experimental implementation based on the prototype's setup in laboratory. It consists of a 3-kW photovoltaic grid-connected inverter, which contributes to a small distortion in voltage at PCC, and practical nonlinear loads connected to PCC. Also, the harmonic-current-injection-model-based time-domain simulations are carried out to prove the potential of the proposed method under various circumstances with different nonlinear loads.