Learning Forum

[mekafime]

Hello everybody,

From a mesh convergence study for a component with a solid element, the values I get in the graph: element length vs strain, the strain is around 0.5 mm/mm, for a ductile material (ASTM A500 gr.B - hollow section square) has an elongation at break value of 23%. What conclusion could be reached based on that value, it seems to me that it is a lot since it should not exceed 2%, although in this forum it was commented that a ductile material has a deformation of 0.5.

[Armin_A]

Hello there,

It is important to differentiate between global and local ductility of materials. The elongation to fracture is a measure of "global" ductility while local strain at fracture is a "local" measure. Your material is steel, so it likely suffers from plastic localization before fracture and the strain field becomes non-uniform; therefore, it is expected that the local strain (shown in your figure) is higher than the elongation to fracture.

Also, please note that ductility is a material-dependent property so 0.5 may not be representative for all materials.

[mekafime]

How do I get that value?, the numerical analysis is not based on a test

[Armin_A]

As far as I know, fracture strain of materials is obtained with tests.

For your particular material, you may be able to find resources online if such tests have been conducted and published by others. 

[mekafime]

In an article found:  Conventional procedures to calibrate constitutive parameters entail the use of standard specimens (such as smooth round bars or flats), in which the stress/strain states are homogenous, allowing for estimation of stress strain directly from the test data (based on specimen cross-section and gage length), and subsequent calibration of the constitutive model from the stress-strain histories. However, these specimens show some form of localization (necking under tension or buckling under compression) at strains in the range of 10–15% (0.1 to 0.15), which is significantly lower than strains at which ductile fracture usually initiates (0–0.75–1.0). Consequently, these specimens cannot be used to accurately calibrate constitutive moover strain ranges that control ductile fracture (0.5–1.0 – Kanvinde and Deierlein [3]). This necessitates the use of non-standard specimens (such as Cylindrically Notched Tension, CNT – see Fig. 3) in conjunction with complementary CFE simulations to back-calculate model parameters that minimize error between the load deformation (rather than stress-strain) histories.