Generalised Cassie–Mayr electric arc furnace models

Abstract

Electric arc furnace (EAF) is one of the largest loads in the power systems. Unfortunately, it is highly non-linear and time varying which causes power quality problems such as harmonics and flicker. Therefore, having an accurate EAF model is necessary. Cassie model is one of the most utilised EAF models in the related fields. However, actual data from electrical system of Mobarakeh Steel Company in Isfahan/Iran show that this model is unable to take into account some important quantities such as the active power and harmonics. Hence, as the first step in this study, different Cassie–Mayr model variants (include the Cassie model) are investigated and the best variant is attained. A novel procedure using large number of recorded actual data is utilised for the models assessment. In the second step, two generalised types of the original Cassie–Mayr model are proposed. Both the generalised types are more accurate than the best-selected Cassie–Mayr variant. All the proposed models have time-varying parameters. Their time-varying nature is studied and by analysing the time series, the proper auto regressive moving average models are attained for every parameter.