Determining failure with stress concentrations present 2

[Stezza]

Hi Guys,

What methods do people use to determine if a part has failed (i.e broken into two bits) when doing an linear-elastic analysis.

When no stress concentration are present this is straight forward as you compare maximum model stress with the ultimate tensile strength of the material (Ftu).

However, when part has geometric stress concentrations at the critical xsection in reality there will be yeilding and stress redistribution which is not picked up by the linear elastic model.

Doe anyone have rules of thumb that they use in this situation?

A specific example that may clarify the question is: If a plate with a central hole has an axial load applied; then by hand calculations the ultimate load that can be carried by the plate is the ultimate tensile strength of the material times the nett xsectional area over the hole (P_ult = Ftu x Anet). Now when you model this in FE you obviously have the effect of the stress concentration of 3 due to the hole showing up giving you a stress at the edge of the hole 3 times Ftu when you apply the load P_ult calculated above.

Also imagine the above example in bending

 

[corus]

You should look at Design rules rather than consider 'rules of thumb'. These design rules such as BS 5500, ASME etc. give rules for classifying stresses and determining their limits. Although primarily developed for hand calculations they can be used with results from FE models with some additional analysis of the results. Some codes enable these features, such as the equivalent linear stress through a thickness, so that the stresses can easily be interpreted against the desgin codes for your particular structure.

corus

 

[hacksaw]

stress concentrations are handled by design factors, the greater concern is use of ultimate strength as your design criteria.

just about any text dealing with design will handle your question, good luck

 

[jetmaker]

Stezza,

Because you are running this model as linear-elastic, the influence of the notch is really not valid once the stresses at the notch exceed yield; hence, neither are your results in this region. If the geometry is rather simple, then I would suggest you take the nominal stress in this region, and back out a load by hand. Then do a notch effect analysis by hand to figure out the theoretical stress. Apply appropriate safety factors and regulatory design criteria to determine if you still retain positive margins of safety.

The other option of course is to run this as a plastic analysis with respective non-linear material properties.

Regards,

jetmaker

 

[corus]

All of the design codes class stresses at a stress concentration as a peak stress. The limit for peak stresses is determined by fatigue limits, and not by the UTS or yield. It is normal practice to use the linearly elastic calculated stress (range) to assess stresses against fatigue design curves.

corus

 

[jetmaker]

Corus,

Although I agree with what you said in your last post, one must be careful with stress concentrations at ultimate load also. In civil engineering where the materials may be very ductile, and large safety factors apply, ignoring the effect of stress concentrations at ultimate load may be acceptable. Where I work, stress concentrations are analyzed at ultimate load because of the low margins of safety, the less ductile nature of the material, and sometimes the proportion of the stress concentration relative to the gross dimensions.

Originally, I was taught that stress concentrations were a fatigue and damage tolerance issue, not ultimate. However, I have since learned different.

There is information avaialable on the web if one searches under "notch effect".

Regards,

jetmaker

 

[Stezza]

Thanks for the replies.

As I am comming from an Aerospace perspective I am looking at the failure condition where the part fails by breaking into two bits. I had been thinking of this failure in terms of the entire cross section reaching the ultimate stress but as Jetmaker pointed out there is a possibility that the sress concentration causes premature failure.

From my reading of a number of NACA technical reports the effect of stress concentrations varies depending on material ductility in two ways (1) ductile materials show strengthing as plastic flow eliminates the initial stress concentration (2) non ductile materials show weakening due to non uniform stresses and strains at the notched region. (Ref NACA-TN-2433 Page 10). Thus you need to know if your material is strengthed or weakended by notches when considering failure.

Also, if anyone has access to an electronic copy of

NACA-TN-1830 "Tension properties of aluminium alloys in the prescence of stress raisers"

and

NACA-TN-1974 "Effect of open circular holes on tensile strength and elongation of sheet specimens of some aluminium alloys"

Could they please post a link?

I will be repeating this question in the Aerospace Engineering forum.

Thanks in advance.

 

[jetmaker]

Stezza,

If you find these papers, I'd also be interested in obtaining copies.

I have a library of such information at home, and will look, but having just moved, everything still is in boxes. I'll let you know if I find any related materials.

Regards,

jetmaker