Statistical analysis of electric-arc-furnace parameter variations

During scrap melting, electric-arc-furnace parameters such as voltage and current, real and reactive power, resistance and reactance etc. exhibit significant variations about longer term mean values. The parameters can be considered to be random variables. The statistical properties of the parameter variations have been evaluated in this paper using data recorded on two 65 MVA furnaces. It is shown that the random variations of a broad range of parameters can be considered to be weakly stationary, but that the variations generally do not follow a Gaussian probability distribution. A significant correlation was found between the resistance and reactance variations per phase. The measured data was used to develop a strategy whereby the mean and standard deviation of the random variations could be measured online, with good accuracy, using conventionally available data-acquisition equipment. It is shown that an improperly chosen sampling strategy can yield quite misleading results.