A predicting model for properties of steel using the industrial big data based on machine learning AITranslate

Basic Information:

DOI：https://doi.org/10.1016/j.commatsci.2018.12.056

Chinese Library Classification Number:

Citation Information:

Highlights • Multiple properties of steel were predicted by nonlinear programming-based model. • One model could predict multiple properties simultaneously. • The potential boundaries of three properties were calculated. Extracting the valuable information about the connections between the overall properties and the related factors from the industrial big data of materials is of significant interest to the materials engineering. At present, most data-driven approaches focus on building a relation model for a single property of the materials, where it may ignore the restrictive boundaries of other properties. In this paper, we propose a machine-learning-based method using nonlinear programming for multiple properties of the materials, and solve the problem by using the Interior Point Algorithm. The key idea is to take the mapping functions corresponding to the properties of the materials as the constraints of the nonlinear programming problem, thus it is capable of processing the restrictions of these properties. Moreover, with our method, the possible boundaries of these properties under certain conditions can be calculated. Experiments results on steel production data demonstrate the rationality and reliability of the proposed method. Graphical abstract Download : Download high-res image (144KB) Download : Download full-size image