Predicting the Electricity Consumption and the Exergetic Efficiency of a Submerged Arc Furnace with Raw Materials using an Artificial Neural Network

Abstract

The problem of higher electricity consumption and lower exergy efficiency in the submerged arc furnace process of the silicon industry needs to be urgently solved. However, various raw materials play important roles in the electricity consumption and exergy efficiency of a submerged arc furnace during silicon production. An artificial neural network (ANN) method was used to model the final strain of the electricity consumption and exergy efficiency with varying silica, coke, coal and electrode. The measured strain versus predicted strain by the model was compared using the R 2 coefficient. The results showed that the exergy efficiency and the electricity consumption values of the testing data are R 2= 0.9918 and R 2 = 0.9896, respectively, in a very short time with low error levels. They clearly indicate the adequacy of the model proposed for prediction of the exergy efficiency and the electricity consumption with different raw materials in the mixture of carbonaceous raw materials in the furnace. Additionally, there is good agreement between the actual and predicted values. Therefore, this developed ANN model is useful to guide the decision about the use of raw materials in silicon production under the condition of lower electricity consumption and higher exergy efficiency.