Premium
Dislocation Density–Based Model and Stacked Auto‐Encoder Model for Ti‐55511 Alloy with Basket‐Weave Microstructures Deformed in α + β Region
Author(s) -
Pang Guo-Dong,
Lin Yong-Cheng,
Qiu Yu-Liang,
Jiang Yu-Qiang,
Xiao Yi-Wei,
Chen Ming-Song
Publication year - 2021
Publication title -
advanced engineering materials
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.938
H-Index - 114
eISSN - 1527-2648
pISSN - 1438-1656
DOI - 10.1002/adem.202001307
Subject(s) - materials science , flow stress , microstructure , dislocation , softening , alloy , deformation (meteorology) , hardening (computing) , composite material , lamellar structure , strain hardening exponent , flow (mathematics) , metallurgy , mechanics , layer (electronics) , physics
The flow behaviors of Ti‐55511 alloy with basket‐weave microstructures are investigated during the hot compression in α + β region. It is observed that the flow behaviors are visibly influenced by the deformation temperature and strain rates. The primary softening mechanisms are the dynamic recovery of β grains and the spheroidization of lamellar α phases. Meanwhile, a dislocation density–based model and a stacked auto‐encoder (SAE) model are built to reveal the flow behaviors of the studied alloy. The relationship between the evolution of dislocation density and the hardening/softening mechanisms are considered in the dislocation density–based model, with the correlation coefficient being 0.9957. The structures of the established SAE model based on the intelligence algorithm are confirmed layer by layer. The SAE model has a high prediction accuracy when the number of hidden layers is 3, and the nodes of three hidden layers are 15, 40, and 35, respectively. It can demonstrate the nonlinear relationship between the flow stress and deformation parameters.