On Wet Film Thickness Measurement System In The Cold Rolled Galvanizing Based On EMD And SVM Algorithm

With the vigorous development of the building, household appliances, automobile industry,China’s galvanized steel production industry is facing unprecedented opportunities and hugemarket. In the production of cold-rolled strip steel hot dip galvanized, coating thickness is animportant subscriber index. If the coating is too thick it will be a waste of raw materials andincrease production cost; or coating is too thin it will affect the corrosion resistance of theproduct, which is a certain extent affect the life of the product. Accordingly, accuratemeasurement of the film thickness and the predictive control has an important significance in theprocess of production of steel sheet coated.A cold-rolled steel galvanized line has introduced a set of NDC wet film thicknessmeasurement system which is capable of accurate measurement data of the wet film, but thecurrent problems is that it can not save the data of the wet film, so the data of the wet film cannot compare with the data of dry film, there is no practical meaning. To solve this problem, theoriginal system has been improved, the OPC (OLE for Process Control) client software isdesigned and developed on the basis of existing hardware and software, the formation of the newNDC-OPC wet film thickness measurement system and the wet film data is stored locally. Inorder to achieve the purpose of predictive control wet film data, firstly discussed the supportvector machine (SVM) regression algorithm, and directed against the defects in the traditionalSVM algorithm, used the Empirical Mode Decomposition(EMD) algorithm to improve the SVMalgorithm applied wet film data values forecasts. Then this paper describe in detail that improvedthe SVM algorithm based on fuzzy information granulation method, and applied to a wet filmdata fitting and forecasting. The experimental results show that the algorithm is feasible.Mathematical modeling of the relationship between wet film thickness of acquisition of thecoating system and the dry film thickness from quality inspection center. To find the relationshipbetween the quality parameters and control parameters from the actual production data, and then,from wet film data has been predicted by the above algorithm and wet and dry film mathematicalmodel to predict the dry film data, provide objective guidance for the control of the productionprocess parameters have important instructive.