Research on Intellectual Prediction for Permeability Index of Blast Furnace

Permeability index of BF is an important monitoring parameter in operation. Proper trend prediction of the permeability index is vital for a good operator. Support Vector Machines combined with the Wavelet Analysis is adopted to build the forecasting model. Four historic values of permeability index are decomposed by Wavelet via seven levels, based on eight Wavelet decomposition combined with operational parameters, using Least Square Support Vector Machines method (LS-SVM), eight sub-models are built. Predicting component are reconstructed to gain the forecast. The detail of modeling, validation and results analysis are presented.