Von Mises Vs Max Principle 2

[TrapperJohn]

Given the legal design requiremnt: The Stress level, under load condition, at any point in the structure shall be limited to a level that provides a safety factor of 3 against permanent deformation.

Whould you use Von Mises or Max Principle in your FEA modle?

 

[johnhors]

TrapperJohn

First point, there is no such thing as "Principle Stress".

Von Mises is a theoretical measure of stress used to estimate yield failure criteria in ductile materials and is also popular in fatigue strength calculations (where it is signed positive or negative according to the dominant Principal stress), whilst Principal stress is a more "real" and directly measurable stress.

Since you have a legal design requirement, then you would be best advised to cover yourself using either stress values.

 

[GBor]

I'm assuming you are using isotropic materials (metal). If so, I usually try to use Von Mises because of the stress combination nature that is considered with this criterion. The metallurgists of the world usually try to drive back to a Max Principle stress because it fits their crack propogation analysis methods. It may help to know what you are designing (more specifically).

 

[Drej]

There is no such thing as "Max Principle". For a static loadcase I would always consider yield against Von Mises stress. Principal stresses are *generally* used in the evaluation of fatigue or fracture based loading (against some value based on a portion of the static yield or the UTS - Goodman approach. Many others are used.).

Where does this requirement come from (ASME or similar)? What are the loads? What is the structure? Which industry (Civil, Nuclear, Marine...) is the structure designed for? Etc.




------------
See faq569-1083 for details on how to make best use of Eng-Tips.com

 

[TrapperJohn]

All

Thanks for the replies and sorry for the spelling. This is a general question for our engineers in the design of aircraft support equipment. Currently some our engineers use Von Mises while other’s use the max/min principal stress. Clearly we all should be designing the same way. Since the requirement to design to is as stated in my 1st post, my question is based on the interpretation of the stated requirement. Could you interpret “The Stress level, under load condition, at any point in the structure shall be limited to a level that provides a safety factor of 3 against permanent deformation” to mean Yield Stress of the material must be greater then 3 times the calculated Von Mises Stress? Or do I need to use the max/min principal stress? Follow up question: how would you define ductile material? Is there a maximum % elongation or some other material property which could define weather a material would be ductile enough to use Von Mises?

 

[rb1957]

i'd interpret the requirement as the maximum value for stress (von mises or principal depending on your preference) to be less than 1/3 of yield, at limit load (which i'd interpret as the loads they've specified). there shouldn't be that much difference between vM or principal, but if there is either use the more conservative one, or the more convenient one (!!) and state which one you've used. i'd doubt that, if you met one of these criteria, that the structure is dangerous; and it'd only be "legal bitching" to say you were diliquent on the requirements 'cause a different interpretation of stress missed the mark by a small margin (in this case you could say that we stated how we met the critieria and it's up to the customer to know the difference, if it matters that much to him).

hope this helps,

 

[gio1]

Given possible legal implications I would use the safest criteria and clearly state it

 

[PhilipFry]

That would be Max Shear criteria then, wouldn't it?

 

[prost]

Pardon my ignorance, but why can't you define maximum
principal stress? If there are 3 principal stresses, and one is larger than the other 2, doesn't that make it the maximum principal stress? (I assume the dispute wasn't over the spelling of 'principal' but over the concept of max. principal stress.)

Yield criteria are a completely different discussion. The main reason in my experience for engineers using different yield criteria (of which there are many: max. principal stress, von Mises yield, max. shear stress, etc.) to define failure is that one criterion seems to have worked better in the past than others. Assuming that this experience is based on reliable test data, it is hard to argue against a particular choice of yield criteria. The selection of Yield Criterion seems to be mainly experience based.

Since the criterion given "The Stress level, under load condition, at any point in the structure shall be limited to a level that provides a safety factor of 3 against permanent deformation" seems to be a purely static strength criterion, I agree with rb1957 that this means limit load is 1/3rd of the yield stress for the material.

If the object being designed undergoes fatigue loading, then a completely different method must be used to ensure design safety (for instance, crack initiation or crack propagation).

I am interested in finding out TrapperJohn would propose to 'unify' the selection of Yield criterion (after having made statement "Clearly we should all be designing the same way") given the huge amount institutional bias (prior experience, design tool documentation, informal "best practices" that are defined in most organizations, etc.) in the selection of Yield criterion for design optimization. Define a test or a series of tests that could be used to test criteria?

 

[prost]

Isn't the von Mises criterion a way to predict failure in a 3D stress state by using the uniaxial stress strain curves? Stress strain curves are almost always derived from simple, uniaxial stress (or strain) tests. From this test, you get such key material parameters as (among others) Young's modulus (stiffness), 0.2% yield stress, ultimate stress, ultimate strain, and maximum elongation. Since your material data is for a uniaxial stress condition, but your object's stress condition is most likely three dimensional, then von Mises stress (which gives you a measure of the three dimensional stress state) is the stress you use to compare to the yield stress (if this is your design limit) derived from the uniaxial stress curves and then to decide how close to the design limit you are.

 

[rb1957]

von Mises and max principal are both contenders for failure criteria ... really they both describe any critieria (failure or yield). i'd only apply them to plates. In aircraft design with stiffened, thin gauge sheet panels we have other strength criteria (crippling and diagonal tension).

 

[rb1957]

sychronised posting ? ...

prost you're perfectly correct ... max principal and vn Mises are both accepted ways of describing the complex stress state of a structure in terms that are comparable to uni-axial test (strength) data.

 

[prost]

Would it be worthwhile for everyone to describe what static strength failure criterion they use and for what situations? I know I'd be interested in such a list. If you also described the rationale for your choice, that would also be useful to the Forum!

(myself, I rarely do static strength analysis, being more often concerned with fatigue, fracture and damage tolerance).

 

[TrapperJohn]

Would it be worthwhile for everyone to describe what static strength failure criterion they use and for what situations?

Sounds like a good idea!

 

[rb1957]

i apologise in advance for this post, but i just can't resist (i did try) ...

i think everyone listing their static strength criteria and how they apply them would be next to pointless.

first, we represent many different industries which naturally focus on different things.

second, if you're working in an industry, you should know common practice

third, if you don't know, then these forums (sorry, i refuse to say fora) really aren't the place to educate your self (on such a general topic).

i'll go now.

 

[GregLocock]

Well said rb. However: to get the ball rolling, I ignore stresses completely when using FEA, for most jobs that I do.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[vorwald]

From memory, I believe that in the NASTRAN theory manual there is a discussion of failure criteria (or margin of safety) that illustrates the different criteria on a stress plot. I'm too lazy to walk upstairs to check the manuals right now.

Similar informat can be found with a google search, such as "von mises stress shear failure criteria"

http://www.efunda.com/formulae/solid_mechanics/failure_criteria/failure_criteria_ductile.cfm

http://courses.washington.edu/me354a/strength_theories.pdf

http://www.mech.uwa.edu.au/DANotes/SSS/failure/theories.html

In the past, for metals, I have generally used Von Mises to identify the highly stressed areas, and then looked at normal and shear in that area.

Also need to check buckling, cripling, etc.

Also, I've experienced more difficulties in getting agreement on the which failure criteria to use with composites.

I would like to see a list of failure criteria also. Everytime I write a stress report, it is a topic of discussion.

 

[GBor]

For composites, there isn't one "all encompassing" stress criteria unless you have performed a lot of material characterization testing and are very comfortable with your test results. After that, I usually use max strain and check each of the directional strains.

 

[vorwald]

From memory, I believe Tsi-Wu indicates failure prior to max strain, see NASTRAN theory regarding margin of safety. So that is part of the issue. And, of course, we are talking about limina strains, not laminate strains. The main problem I have is lack of coupon data for the allowable off axis strains and also allowable. When asked/requested/directed, I've tried to estimate reasonable values for these allowables (from coupon / theory) and I tend to find that designed structures don't satisfy reasonable values for off-axis or shear allowables. Just indicates that criteria wasn't used during the design. Anyway, I haven't done a composite analysis for a couple of years, and don't have any planned in the foresable future. Previously, I have obtined I did get good correlation with coupon values for strain distribution around a hole.

 

[cbrn]

Hi,
there is not a "better" or "worse" criteria to combine stresses. There are plenty, which apply for ductile materials, or for fragile materials, or for laminar composites, and so on.
Of course, if you trace the sigma-tau plot of all these criteria, you may find that some are more "conservative" than others, BUT the real point is that you may have to choose the most appropriate for YOUR kind of material (see Drej's questions...).
Another point to pay attention to is: do you have to respect a NORM? If so, you don't have any choice: you MUST use the specified criteria. For example, if you decide that you will design a pressure vessel according to ASME, because you think ASME is recognized "de facto" worldwide, you will have to use Trescà-Guest. In the same scenario, if you decide to use European "PED" with the appropriate EN norm, you will have to use Von Mises.

Regards