A Probabilistic Three-Phase Time Domain Electric Arc Furnace Model based on analytical method

An electric arc furnace (EAF) is known as nonlinear and time variant load that causes power quality (PQ) problems such as, current, voltage and current harmonics, voltage flicker, frequency changes in power system. One of the most important problems to study the EAF behavior is the choice of a suitable model for this load. Hence, in this paper, a probabilistic three-phase model is proposed based on recovered hidden Markov model (RHMM) in time domain. To recover the HMM , the coupling factor is proposed. This factor estimates the past observations and considers the effects of all observations in different states. Regarding to the intense fluctuations of various parameters of EAF, this factor can improve the EAF model in different operating stages. This subject causes that the proposed model is closed to the actual model. To train the RHMM, actual measured samples are used. Likewise, different parameters of EAF' power system as, flexible cables, electrode, busbar are exactly considering to achieve an accurate model. Comparing the results of experimental and proposed model indicates the accuracy of the proposed model.