Solid Finite Element vs Through-thickness stress close to yield point 1

[EngineerMickeyMouse]

Hi gents!

I would like to ask you about your engineering opinion on following FEA case.

Given is thick plate (40mm thickness) with dimensions in plan of 500x500mm. It is modelled by solid finite elements with edge of 10mm each.
At some small area of this plate, in effect of loading conditions, there is cubic area 40x40x40mm for which mean von Mises stress is equal to yield strength of the ductile steel material plate is made from. Other areas are well below yield strength.
Note, it is through-thickness phenomenon.

Question is, how would you interpret the results and why, will the plate take it or not?

 

[rb1957]

it sounds like your plate has yielded, and maybe needs a non-linear analysis.

another day in paradise, or is paradise one day closer ?

 

[EngineerMickeyMouse]

That is what I think too. Literature and various codes are giving following acceptance criteria for nonlinear problem, it is peak strain allowable value of 5% for bilinear stress-strain material curve. Now one may ask if this 5% value should be found at one FE node value or averaged among some cubic/area? If one value, how to exclude any singularities from this, perhaps it is true that in its definition - strain values cannot have singularities, do you agree, rb1957?

 

[rb1957]

if you exceed yield through the thickness for a sizeable area, you've probably yielded the plate. Results from a linear model aren't usable.

another day in paradise, or is paradise one day closer ?

 

[GJD56]

Keep in mind that the yield stress is traditionally set at 0.02% strain. This is not the elastic limit of the material. So, if you are at or above the YS of the material, you have probably already exceeded the elastic limit of the material.

Secondly, if you are above YS, the elastic model is still useful in that you can alter the design and compare the resulting stresses without having to go full nonlinear. Besides, if you are using common plate, the nonlinear characteristics of this material is quite complex making the analysis perhaps even less useful. I would work to reduce the stress by changing geometry if possible to get it below the YS point.

 

[CooperJerWI]

Given that the plate is 40mm thick and the length of the long side of the material is 500mm you may consider running the FEA as a shell or thin plate. The general rule I've read indicates a plate can be considered "thin" when the thickness is 1/10 or less of the long dimension on the plate. I generally use 1/15 to 1/20 with the FEA software I use. The results will give you another data point to consider how accurate your brick analysis is. In addition, there should be a precision analysis feature in your software. Follow up on that tool to determine where 10% precision is. ANSYS published some documentation some time ago indicating they would recommend a stress precision of 10% or less for accurate results.

 

[SacovanLoon]

As it is through thickness stress, the material will yield in shear. A hole will be punched.

Saco van Loon

http://www 2st

nl

 

[VN1981]

OP, what is the loading condition the plate is subjected to? If its transverse loading, are you applying a pointed load on a node? If so, as you have mentioned, the local peak stresses could be due to local high force concentration. Also what are the fixity or support conditions for the plate?

This is one check I can think of. Please see the variation of Von Mises stresses from loading point to some short distance. If the variation is huge, then the anomaly could be due to loading effect i.e. if the distance to which stresses are below yield is really short.

"At some small area of this plate, in effect of loading conditions, there is cubic area 40x40x40mm for which mean von Mises stress is equal to yield strength of the ductile steel material plate is made from."

"Note, it is through-thickness phenomenon."

You are quoting both Von Mises as well as Through-thickness. It is possible that major component in Von Mises is coming from Through Thickness (is the through thickness stress normal or shear components?)