A Fig-Based Method for Prediction Alumina Concentration

Existing prediction model can not be established accurately as a result of there is often a lot of redundant information in observed values of alumina concentration. A prediction method based on fuzzy information granulation for alumina concentration is proposed to solve above problem. In the proposed approach, theory of fuzzy information granulation was used to granulate time-series data of alumina cell. Granulated data can not only reflect the characteristics of original but also reduce redundant information. Support vector machine was employed as predictor. The experimental results using real data of 170KA operating aluminum cell from a factory demonstrate the efficiency of the designed method and the viability of the technique.