TY - JOUR
AU - Bin Cao
AU - Shuang Yang
AU - Ankang Sun
AU - Ziqiang Dong
AU - Tong-Yi Zhang
TI - Domain knowledge-guided interpretive machine learning: formula discovery for the oxidation behavior of ferritic-martensitic steels in supercritical water
JO - Journal of Materials Informatics
PY - 2022
VL - 2
IS - 2
SP -
EP - 4
SN -
AB - A formula with high generalization and accurate prediction power is most desirable to science, technology, and engineering. In addition to human beings, Artificial Intelligence (AI) algorithms show a great promise to discover formulas. In this paper, we propose a domain knowledge-guided interpretive Machine Learning (ML) strategy and demonstrate it by studying the oxidation behaviours of Ferritic-Martensitic (F-M) steels in Supercritical Water (SCW). The oxidation Cr equivalent is, for the first time, proposed in the present work based on a generalized Arrhenius oxidation formula suggested by our domain knowledge and interpretive machine learning algorithms. Tree-Classifier for Linear Regression (TCLR, Open Source) algorithm is, also for the first time, developed to materialize the formula with collected data. TCLR effectively captures the linear correlation between compositions, testing environments and activation energy, time exponents from the data. The Sure Independence Screening and Sparsifying Operator (SISSO) algorithm finally assembles the information derived from TCLR, resulting in a general formula. The general formula with determined parameters has the power to predict, quantitatively and accurately, the oxidation behaviours of F-M steels with multiple alloying elements exposed to various SCW environments, thereby providing guidance to the design of anti-oxidation steels and hence promoting the development of power plants with improved safety. The present work demonstrates the power of domain knowledge-guided interpretive ML with respect to the data-driven discovery of physics-informed formulas and to the acceleration of materials informatics development.
KW - Domain knowledge
KW - interpretive
KW - oxidation Cr equivalent
KW - tree classifier for linear regression (TCLR)
KW - general formula
DO - 10.20517/jmi.2022.04
UR - http://dx.doi.org/10.20517/jmi.2022.04