Investigation On Soft Sensor Technique Of Sinter Mixture Moisture

The sinter is a primary method for improving the using rate of lean ore, because the rich ore is lacking than the lean ore in my country. Now the sinter ore is mostly part of the blast furnace charging, it is about 90 percent. The sinter process directly influence the output and the quality of the agglomeration, so as to the smelting.The sinter is a complicated physical chemistry process. It includes the vaporizing and condensation of moisture, the decompounding of solid materiel, oxidation deoxidize reaction, reaction and the melting and condensation of chemistry matter etc. The process makes the sinter mixture that is consist of powdery ore and return ore powder heated highly by the bake machine and agglomerated under deficiently dissolved condition. The moisture of mixture in this process is an important factor that affects sinter process. The appropriate moisture felt powdery ore and returned ore powder. It makes moisture adequately infiltrating into the charging. Under this condition it has minimal stack -proportion as good as maximal cubage, best permeability and being prone to globularity.However, It is difficult to measure moisture directly as complexity and uncertainty of the sinter system. Therefore, this paper brings forward the sinter mixture moisture measure system based on permeability soft sensor technique model, and also, though study the technical and process of sinter that use mixture moisture as assistant variable and the modeling method by soft sensor technique. The paper adopts three modeling: neural network, fuzzy control and fuzzy and neural network. Being based modeling means and experiment analysis shown that subtractive clustering fuzzy neural network model has better astringency, upper forecast precision, lesser testing error and adapts to engineering application by comparing.The experiment shows that the method of mixture moisture soft sensor technique not only effective but also has upper measure precision.