A neural-network-based data-driven nonlinear model on time- and frequency-domain voltage-current characterization for power-quality study

An accurate model of nonlinear load is important for the evaluation of power quality (PQ). Among different PQ disturbance sources, alternating current electric arc furnace (AC EAF) is one of most complicated and serious loads. To provide effective operation prediction of AC EAF, a data-driven modeling approach based on time- and frequency-domain voltage-current (v-i) characterization is proposed in this paper. With the prediction of the proposed model in the time domain, the dynamic and short-term behavior of AC EAF can be observed. And the quasistationary and long-term features of AC EAF would be extracted via the frequency-domain phase of the proposed model. From the comparison on the field measurement data, the performance of the proposed model can be verified in the applications of PQ studies.