Detecting Defects of Steel Slabs Using Symbolic Regression

The quality of products of heavy industries plays an important role because of further usage of such products, e.g. bad quality of steel ingots can lead to a poor quality of metal plates and following wastrels in such processes, where these metal plates are consumed. Of course, single and relatively small mistake at the beginning of a complex process of product manufacturing can lead to great finance losses. This article describes a method of defects detection and quality prediction of steel slabs, which is based on soft-computing methods. The proposed method helps us to identify possible defects of slabs still in the process of their manufacturing. Experiment with real data illustrates applicability of the method.