Difference between Von Mises and maximum Principal stress

[Bilal91]

What is the difference between Von Mises and maximum Principal stress ?

I have a Brittle material (WC-10% Co), and I would like to see when it fractures. I have Transverse Rupture Strength value (3700 MPa).

Can I compare the Max Priciple Stress with TRS with ignoring Von Mises ??

I am using Ansys Workbench 15.0 to apply a radial force on a tip of an End mill tool to check when it fractures.

 

[stressebookllc]

For brittle materials, max principal is more appropriate. Von Mises is more suitable for ductile materials like steel.

http://www.stressebook.com

Stressing Stresslessly!

 

[Bilal91]

Attached is the first picture of the results I got. (Maximum Principle Stress).


Note: TRS value is 3078 MPa while the Von Mises is 17350 MPa !!!!!!!! As you see, it's a huge difference !!!
So, can I compare the Max Principle Stress with the TRS and ignore the Von Mises ??

 

[Bilal91]

Here is picture 2 (Von Mises)
I applied a force of 3500 N.

 

[glass99]

Three thoughts:

1. max principal stress basically ignores shear stresses. It is basically the maximum tensile stress vector.
2. We use principal stress for all of our glass analysis because it is brittle. But brittle is a relative term. Is there any yielding in your material?
3. Your von Mises hotspot is a hyper localized. Are you sure there is nothing funky going on locally like a distorted element or something? There is usually not such a big difference between vM and principal.

 

[Bilal91]

Attached is a picture of the mesh and position of the group of nodal forces applied as a big nodal force (Not to have a concentrated stress if I apply 1 single nodal force).

Second, I don't have a yield value in the data sheet. Instead I have the TRS value and K1C and E-Modulus..

 

[rb1957]

"max principal ignores shear stresses" ... no, principal stress combine normal and shear stresses to give the equivalent normal stress. there is also one or two other principal stresses acting at the same time (the element is not in uni-axial stress) and von mises combines these into a single failure critieria.

a large difference between vM and principal shows that there are large principal stresses on the other axes.

personally I'd use whichever (vM or principal) was larger.

"Transverse Rupture Strength" is an unusual name for Ftu ... how is it determined ?
"Transverse" implies across the thickness, or the grain, ... are the in-plane strengths higher ?

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

 

[Andries]

Some nit picking: The principal stress is a real stress (tensor) while the Von Mises stress is a scalar quantity based on the square root of the sum of the squares of the stress tensor components.

Andries

 

[rb1957]

yes, principal stresses are real, and are often used was a failure index (compared to Ftu).

von Mises is a construct, based on principal stresses, that is a failure theory that combines the three principal stresses into one number, to compare to Ftu.



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

 

[stressebookllc]

Agreed with RB1957, Von Mises will not directly show you whether the stress state there is tensile or compressive or shear (if you look at the deformations you can probably tell though). You simply cannot have a huge Von Mises with low contributing normal and shear stresses because Von Mises stress is an equation that uses different stress values.

Check all the other stress values there too, max, min and shear stresses. You are using solid tet elements, hope you are not using linear elements as those are stiffer than parabolic tet elements and tend to be on the conservative side.

http://www.stressebook.com

Stressing Stresslessly!

 

[Bilal91]

Hey all,

Thank you for your responses.
Yes, I am using tet elements for my mesh.

Here is a picture as a print screen from a book says that:
For Brittle materials: If the Maximum Principle Stress reaches the Fracture Strength (TRS) of the material, it will fail.
So, I'll use it simply.

 

[adfergusson]

I'm going to add my two cents with a perspective of someone who selects tools for machining parts (and who uses FEA):

For a start; tools fail due to wear/fatigue/damage accumulation, not simple one off peak loads (unless the person selecting and using the tooling doesn't know what they're doing)

Anyway, where are your forces coming from? Applying a force to a set of nodes seems arbitrary. The force that a tool experiences is dependent on the cutting surface geometry, the material being machined, the cutting strategy, the feeds and speeds involved plus a whole bunch of other parameters for varying significance. See the attached pic of a cutting force vs rake angle for PA 4/6 for example (from Yatish Patel's PhD thesis; The Machining of Polymers, Imperial College, 2009) The way that chips form during cutting is a function of the material being cut along with the geometry of the tool as well as the aforementioned speed and feed rates...lets take an analogy of grabbing a knife and trying to shave/scrape off a layer of material from, lets say, a bar of soap. The angle at which you hold the blade to cut will have a huge effect on the amount of force/energy to cut through a given amount of material and also whether the soap you cut comes off in a continuous sliver or breaks up into smaller slivers. Consequently you can't simply compare the apparent failure load of tools with different geometries; one may appear stronger than the others but if it results in a higher cutting fore then that would mitigate such a difference.

And the above considerations are just for starters. For example, another important factor when dealing with tooling, especially brittle tooling, is the runout of the tool holding assembly. Runout is a value that described the eccentricity of the tool and there is plenty of literature out there on the effect of runout on tool life.

I don't think that FEM would be the easiest/fastest/cheapest approach for selecting or designing a new tool. If you're doing this for research purposes/a degree then that's great but I expect you'll need an explicit solver (e.g. LS-Dyna for ANSYS), plus an Intel Fortan compiler for implementing appropriate user defined materials, to really get into this.

 

[glass99]

What is the cut off between when you switch from principal to vM stress failure criteria? I assume there is some shade of grey if you have a material such as a hardened steel which has a ultimate 10% more than its yield.

 

[rb1957]

going back to your 10th Jan post ... "Note: TRS value is 3078 MPa while the Von Mises is 17350 MPa !!!!!!!! As you see, it's a huge difference !!!"

ok, what is the corresponding max principal ?

is this a very localised peak stress ? even with a brittle material this could be a problem (with a ductile material you have "local plasticity" as an out). but is it a modelling issue ?? (a point load, a hard constraint, ...)

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