XANSYS Message: 49592 [Go back to message list]
[bookmark on del.icio.us]
No rating yet
Rate item:

Subject: Re: [struct] Bolt Stress Concentrations
Author: Ian Prentice
Date: 2003-04-25 11:28:00

In addition, 'good' bolts are manufactured using a cold rolling process in the
head fillet and threads. This imparts local residual compressive stresses which
offset the tensile stress from loading, so delaying the onset of yield. Further,
the cold rolling process work hardens the bolt locally (depending on the
material) increasing the yield point in the area of peak stress. A simple
elastic analysis will always under predict the strength of a bolt, except for
machined bolts, which should not be used for high strength applications.
Ian
GEAE

-----Original Message-----
Sent: Friday, April 25, 2003 10:49 AM
To: ?

>the model is showing massive stress concentrations around the 'under head'
> radius. This is, of course, to be expected, but would classically indicate a
potential
>failure.
Not necessarily under static loading--go back and read your mechanics of
materials. True stress concentrations cause local yielding which simply
shifts a small portion of the load around a little bit. The strain is
limited because the corner or notch, whatever, is encapsulated in a large
area which remains elastic. Your bolt head, and the thread roots,
experience these concentrations, and a little yielding and redistribution
takes place, but with no more effect than the plastic deformation when
the threads were rolled or the head swaged into shape.

It's different when you get cyclic loading and the concentration
accumulates plastic strain with each cycle. Eventually a small rupture
forms when the strain capacity is exhausted and fatigue eventually occurs
as the crack grows to a critical length. That's why you pre-load
bolts--to reduce the cyclic load component.

For milder stress concentrations, the local stresses 'shake down' to
elastic action, if the local stress doesn't exceed twice the yield, as
calculated assuming purely elastic behavior. The first load cycle
produces first elastic then plastic strains during load application, but
when the load is removed the stress drops off linearly to zero and then
compression, producing a built-in stress not much different that the
stresses built in during cold work. The next load cycle goes linearly
from compression to zero to tension with less plastic strain than before,
because deformation remains elastic for a greater portion of the cycle.
This is called shakedown, and most real structures undergo something of
the sort early in service. There are exceptions, notably for brittle or
deeply flawed structures where plastic deformation is limited. Materials
which strain soften don't shake down in the same way, and plastic strains
may cause unacceptable deformation in precision parts.

Precise assessment requires that the engineer know the exact stress state
at the instant the first load cycle is applied. Each step and operation
during fabrication be analyzed with micrometric precision for plastic
strain accumulation. Then entrails must be examined to determine all
future load cycles and the sequence of application so each increment of
plastic strain is accounted for with a wonderful non-linear analysis.
Unfortunately the entire structure must be converted into test specimens
to provide the material properties to prove satisfactory service had you
actually been able to put the thing into service. Equally unfortunate, if
you find a problem you have to do it all over again. Or you can do it the
easy way by using ANSYS as a tool, rather than an object of worship. (I'm
kidding... everyone got that? Just kidding...exaggerating to make a
point. It's Friday, remember?)

>The Ansys
>help people have assured me that the model is correct, and that this type of
>anomaly should be expected and not taken literally, but this argument is
>quite difficult to take to management!
Put yourself in their position. They arm someone with all the latest
analytical tools and the best he can do is tell them that things which
have always worked don't work. That's why I harp constantly on
understanding the physics of the problem. Putting numbers to a problem
without a physical understanding of the process is like driving
blindfolded--you blunder around and crash into things and you get
nowhere.

Christopher Wright P.E. |"They couldn't hit an elephant at
chrisw@s... | this distance" (last words of Gen.
___________________________| John Sedgwick, Spotsylvania 1864)
http://www.skypoint.com/~chrisw
Posts possibly associated with message #49592AuthorDateScore
49576[struct] Bolt Stress ConcentrationsCooper, Geoff geoff.cooper@2003/04/25 
49579Re: [struct] Bolt Stress ConcentrationsMarco Perucchi2003/04/25 
49581Re: [struct] Bolt Stress ConcentrationsMark Krueger2003/04/25 
49585Re: [struct] Bolt Stress ConcentrationsJohn Ceko2003/04/25 
49590Re: [struct] Bolt Stress ConcentrationsRonnie Thomas2003/04/25 
49591Re: [struct] Bolt Stress ConcentrationsChristopher Wright1969/12/31 
49592Re: [struct] Bolt Stress ConcentrationsIan Prentice2003/04/25 
49594Re: [struct] Bolt Stress ConcentrationsPhilippe Vidori2003/04/25