EFFECT OF MICROSTRUCTURES ON STRAIN HARDENING EXPONENT PREDICTION OF CAST ALUMINUM ALLOY

Abstract

Strain hardening exponent(n) of a material is an important parameter reflecting its hardening property whose determination is of great importance.It has a widely application in material scientific research and engineering fields such as fatigue life prediction,stress-concentration factor calculation,etc..The value of strain hardening exponent varies with their microstructures in cast aluminum alloys,but a few theoretical and experimental investigations have been reported to understand the effects of the microstructural parameters on the strain hardening exponent in these alloys till now. In the present study,the influence of secondary dendrite arm spacing(SADS),aspect ratio and volume fraction of particles on internal stress in aluminum alloys were discussed,and a quantitative prediction of strain hardening exponent was established on the basis of Hollomon and internal stress equations.It shows that the strain hardening exponent represents their hardening ability in relatively large plastic deformation(larger than the upper limit for the no plastic relaxation regime).A new microstructural hardening parameter relatively to strain hardening exponent was defined.Besides,a group of A319 cast aluminum alloys with microstructural heterogeneities were tested.The calculated strain hardening exponents are in agreement with the experimental ones in A319 alloy as well as in some commonly used cast aluminum alloys.For the same grade of alloys,the microstructural hardening parameter and strain hardening exponent are quite sensitive to SDAS and particle aspect ratio while the influence of volume fraction of particles is relatively little.As the values of SDAS and aspect ratio of particles increase,the value of the strain hardening exponent decreases.A liner relationship between microstructural hardening parameter and strain hardening exponent was proposed.For A319 and A356/357 alloys,the optimum correction coefficients are 0.17 and 0.11,respectively,and the mean prediction error of n is only about 10%.