Artificial intelligence model of complicated flow behaviors for Ti-13Nb-13Zr alloy and relevant applications
(1. CQU-UC Joint Co-op Institute, Chongqing University, Chongqing 400044, China;
2. State Key Laboratory of Mechanical Transmission, College of Material Science and Engineering, Chongqing University, Chongqing 400044, China;
3. School of Civil Engineering, Chongqing University, Chongqing 400045, China)
2. State Key Laboratory of Mechanical Transmission, College of Material Science and Engineering, Chongqing University, Chongqing 400044, China;
3. School of Civil Engineering, Chongqing University, Chongqing 400045, China)
Abstract: The comprehensive nonlinear flow behaviors of a ductile alloy play a significant role in the numerical analysis of its forming process. The accurate characterization of as-forged Ti-13Nb-13Zr alloy was conducted by an improved intelligent algorithm, GA-SVR, the combination of genetic algorithm (GA) and support vector regression (SVR). The GA-SVR model learns from a training dataset and then is verified by a test dataset. As for the generalization ability of the solved GA-SVR model, no matter in β phase temperature range or (α+β) phase temperature range, the correlation coefficient R-values are always larger than 0.9999, and the AARE-values are always lower than 0.18%. The solved GA-SVR model accurately tracks the highly-nonlinear flow behaviors of Ti-13Nb-13Zr alloy. The stress-strain data expanded by this model are input into finite element solver, and the computation accuracy is improved.
Key words: Ti-13Nb-13Zr alloy; flow stress; constitutive model; support vector regression; genetic algorithm