Constitutive modelling of artificial neural network for hydrogenated Ti65 alloy based on genetic algorithm optimization
The study conducted isothermal compression tests of Ti 65 alloy samples at different hydrogen contents(unhydrogenated,0.13 wt.%,0.25 wt.%,0.34 wt.%,and 0.43 wt.%hydrogen)in the α+βtwo-phase and βsingle-phase regions at a strain rate range of 0.001 s-1 to investigate the high-temperature deformation behavior of hydrogenated Ti 65 alloys and construct a GA-BP constitutive model for Ti65 alloys that considers hydrogen content,deformation temperature,strain,and strain rate.The model was integrated into the finite element software to simulate the isothermal compression process of hydrogenated Ti65 alloy samples.The results showed that the correlation coefficient and the average relative absolute errors value of the 4-12-12-1 structure GA-BP constitutive model were 0.998 2 and 0.46%,respectively,with high prediction accuracy and generalization ability.The simulation results of isothermal compression indicated that the established GA-BP constitutive model had high simulation accuracy and could be used for analyzing the thermoplastic forming process for locally hydrogenated Ti 65 alloy.
hydrogenationTi65 alloyartificial neural networkgenetic algorithmconstitutive model
NETL
NSTL
万方数据
