Modeling of the laminar cooling process with case-based reasoning

A hybrid modeling approach integrating case_based reasoning (CBR) into the first principles process model is developed for the laminar cooling process,which is difficult to describe with accurate mathematical models due to its complex nature(e.g.,highly nonlinear,time varying,and spatially varying).The past cases most similar to the current operating condition are retrieved from the case base.The model parameters of the finite difference model for the current operating condition are found by reusing the solutions of the retrieved cases.The finite difference model can predict the through_thickness temperature evolutions of the moving strip during the cooling process.Experimental studies based on the industrial data from a steel company show the effectiveness of the proposed modeling approach.