Portal de Periódicos da CAPES

Knowledge extraction and transfer in data-driven fracture mechanics

2021; National Academy of Sciences; Volume: 118; Issue: 23 Linguagem: Inglês

10.1073/pnas.2104765118

1091-6490

Xing Liu, Christos E. Athanasiou, Nitin P. Padture, Brian W. Sheldon, Huajian Gao,

Drilling and Well Engineering

Significance Data-driven approaches have launched a new paradigm in scientific research that is bound to have an impact on all disciplines of science and engineering. However, at this juncture, the exploration of data-driven techniques in the century-old field of fracture mechanics is highly limited, and there are key challenges including accurate and intelligent knowledge extraction and transfer in a data-limited regime. Here, we propose a framework for data-driven knowledge extraction in fracture mechanics with rigorous accuracy assessment which employs active learning for optimizing data usage and for data-driven knowledge transfer that allows efficient treatment of three-dimensional fracture problems based on two-dimensional solutions.