The multi-scale forecast of submerged arc furnace energy consumption base on support vector machine

Due to ferroalloy submerged arc furnace smelting is an extremely complicated chemistry and physics reaction process, the exact forecast of energy consumption related to the stable and efficient operation of submerged arc furnace, introducing a multi-scale energy consumption prediction model based on least squares support vector machine (LS-SVM), first of all, by conducting the wavelet decomposition to the energy consumption sequence; we can get the approximation coefficient and wavelet coefficient according to the specific scale and related scale. Then, we conduct the multi-scale combined forecast by using the coefficient of predictive position. LS-SVM. Eventually, through the wavelet reconstruction, we can calculate the corresponding predictive value of submerged arc furnace energy consumption. We conduct the simulation test combined with the demand data of submerged arc furnace energy consumption in Sinosteel Jilin Ferroalloys Co.,Ltd. The results indicate that the method introduced in this paper has a considerable forecast accuracy and practical value.