A new adaline approach for online voltage components extraction from unbalanced and perturbed power systems

This work presents theoretical studies and practical results obtained with voltage component extraction. The paper is centered on a new method for estimating the direct, inverse and homopolar voltage components from unbalanced and disturbed power systems. We introduce and develop a new decomposition of the voltages in the DQ-space that results in linear expressions explicitly separating AC from DC components. These expressions are learned by Adaline neural networks because of their simplicity and capabilities to approximate linear relationships and to learn online with respect to real-time applications. While learning, the Adalines estimate the amplitude and the phase of the direct, inverse and homopolar voltages of the electrical network and efficiently compensate for the disturbance events, i.e., time-varying nonlinear loads, parameters changes, noise perturbations, fluctuating distortion harmonics. The method was studied and successfully implemented, simulations and experiments under stationary and nonstationary conditions are reported. A demonstrative comparison with other methods is also addressed.