A neural-network-based method of modeling electric arc furnace load for power engineering study

It is known that artificial neural network is a powerful scheme for function learning and modeling nonlinear loads. However, a direct application of artificial neural network for modeling time-varying loads may lead to inaccuracies. This paper presents an accurate neural-network-based method for modeling the highly nonlinear voltage-current characteristic of an ac electric arc furnace (EAF). The neural-network-based model can be effectively used to assess waveform distortions, voltage fluctuations, and performances of reactive power compensation devices associated with the EAF in a power system. Simulation results obtained by using the proposed model are compared with the actual measured data and two other traditional neural network models. It is shown that the proposed method yields favorable performance and can be applied for modeling similar types of nonlinear loads for power engineering studies.