Screw fatigue in bending. 2

[Biggadike]

I have a group of M8x30 High Tensile (12.9) cap head socket bolts which I am using in a vibratory system (in a vehicle). I have accelerometer data and want to calculate the fatigue life of the screw(s). I am missing two things:

1/ The S-N curve or similar for the material.

2/ The stress already existing within the screw before the oscillating bending forces are applied. - I can have a reasonable stab at calculating these but proper data would be good.

I seem to remember rolled threads giving a 1.4 stress raiser which also needs confirming.

Please don't point me in the direction of a bit of internet software, I'll never get it past our security and I've never had much success with them in the past anyway.

Anybody got any real data?

 

[Metalguy]

If you're lucky your screws are used in a rigid connection. In such cases, if you can get the preload high enough the screws won't see more than a little varying stress. If non-rigid, you'll be busy.

 

[Biggadike]

I heard that said before: essentially that if you clamp it down hard enough there can't be a bending moment and it sorts itself out.

My concern with that was that by clamping it down very hard, the initial stress in the screw is raised further making the screw less able to deal with the vibration forces.

Maybe I'm wrong and this is actually the solution.

 

[Dennyd]

Biggadike,

What you want to avoid is alternating the net stress in the bolt from tension to compression. Think of your vibration load as a sine wave that does alternate from tension to compression with an amplitude of 10 lb (or N if you prefer). If the bolt is tightened with an initial tension of 200 lb then the net stress in the bolt will vary from a max of 200+10 = 210 lb (divided by the cross sectional area of course) to a minimum of 200-10 = 190 lb (/area) all in tension. If the bolt alternates between tension and compresion then the fatigue life is dramatically affected. It is also more likely to loosen, which will only aggravate the situation.

- - -Dennyd, P.E.

 

[CoryPad]

[blue]Biggadike[/blue],

It is true that using high fastener preload means there is less capacity for the fastener to take on additional force transferred from the joint. However, there is a geometry effect as well. The bolt doesn't absorb all of the force applied to the joint. If the joint is designed well, then it will absorb little of the additional load. Using properly sized fasteners tightened as much as possible is the correct procedure for fatigue-sensitive joints. I recommend the following for bolted joint education:

faq725-600
faq725-536

http://www.boltscience.com

Regards,

Cory

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

 

[guyguy]

you said your concern is that the clamping shall cause high pretension in the bolts. well, what are the loads (alternating and others) that each bolt sees, and what type of bolts are you using?

I guess you'll find out that the loads are rather small compared with the bolts yield point, and in that case you can preload them to a satisfying value (and that value is a function of the type of bolt you are using)

cheers,

guy

 

[unclesyd]

Simply stated:

"If a screw or bolt is preloaded (tightened) beyond the working load encountered, changing stresses will not affect it"

Elton McBroom: Sports Car Magazine; January 1964

 

[Biggadike]

OK,

The problem I have with all this is the following:

Fatigue failure happens below yield stresses and it is the tensile forces which propagate the cracks. If the normal pre-load is 90% of the yield stress then fluctuating tensile stresses on top of that are certain to cause a fatigue failure at some number (N) of load cycles, especially as the thread profile gives a locally raised stress of 1.4 times nominal. This takes the root of the thread past yield point.

This screw is used in pairs to clamp 4 stainless steel mounting blocks to a stainless steel plate. The block is only 14.5mm thick, the mounting point is 37mm from the clamping face and is subjected to an extreme sequence of vibrations in a six hour test which equate to 25 years of sitting inside a military vehicle.

This being the case, I hope you can see why I'd like to see a fatigue performance curve for the screw material.

That said, the info posted so far has given me enough to get started and I'm now in the process of calculating the levels of stress the screw would see. So THANKS for that..

 

[Metalguy]

Dennyd,
Yes, resistance to fatigue cracking depends largely on avoiding varying loads, esp. avoiding tension-comp-tension, but the applied loads are not added and subtracted per your example if the joint is rigid and the preload is high enough.

 

[Biggadike]

I've been working through the calculations and (because I keep having to do other things) I'm about half way through. I have established that the root of the thread is in a yield state which means that even in cyclic tension the root of the thread will work harden and crack.

I'm using an empirical equation based on Basquin's equation which equates the number of cycles to failure with the change in strain per cycle.

We shall see what it predicts.

 

[Metalguy]

>"I have established that the root of the thread is in a yield state which means that even in cyclic tension the root of the thread will work harden and crack."<

Your conclusion may be wrong. Many items are loaded beyond the "yield strength" and do NOT develop fatigue cracks. Many others do develop cracks and the stresses are well below the YS and the PL.

 

[Biggadike]

Mmm, Maybe,

But I'm struggling to see how a material which work hardens can do anything but just that when cyclicly stressed in the yield state. If it does work harden, it must logically become brittle and crack. This would reduce the cross section and the process would repeat and increase in speed.

I know material science is tough to generalise but what am I missing in this case?

You are certainly right that you don't need to hit the Yp to get fatigue cracks, but having hit the Yp, cracks ahoy!

 

[Metalguy]

Hmm, back to the basics. Forget about trying to relate YS with fatigue cracking. YS *does* relate fairly well to SCC and HE/HIC, but not fatigue.

Also, try to read the books by Bickford on bolting. You will find that many structural bolts are intentionally tightened past their YS w/o problems.

 

[Biggadike]

I'm still not convinced. I'm not sure which basics you mean but my statement still stands - I have a case study sitting next to me where what I'm describing lead to failure so I'm not going to be so easily deterred. Also, Stress corrosion cracking (SCC as you say) is primarily related to tensile stress (often through process stresses) and corrosion at grain boundaries in materials which show no overriding alternative corrosion state (i.e. copper, aluminium). I see no relationship between SCC and yield strength except the stresses which working the material may give.

I haven't read Bickford and in the next few hours that really isn't an option but if you'd like to summarise what he says then I'd be interested to hear it.

 

[Metalguy]

SCC does have a fairly close relationship with the YS of a material, but your concern is fatigue. Fatigue crack resistance (resistance to cracking initiation) depends on many things, primarily the R ratio (ratio of max/min stresses). *Correctly* preloaded fasteners in rigid connections keep the R ratio favorable--there is little change in the stresses on the fastener in service.

 

[boo1]

agree w metalguy

 

[Biggadike]

I agreed with him too although my post didn't appear.

I am following those good design principals and my calculations give a correspondingly long life expectancy for my screws. All good stuff.

I am interested though in the idea that moving the screw into the yield area doesn't affect the fatigue life. Is that true? Or is it that experience shows that, when using good design practice, fatigue life is so long that it makes no difference.

 

[Metalguy]

Here's another way to look at things. It is generally assumed that as-welded welds and HAZ's have residual welding stresses approx. equal to the YS. How many billions of such welds are in service, including varying-stress service?

Forget about the YS and concentrate on the important factors that *do* affect fatigue resistance.

 

[unclesyd]

Take a look at the following website.
Checkout the bar and rod database.

http://www.steel.org/autosteel/bar_rod/index.htm

 

[Biggadike]

Funny you should mention welds...

The reason for me putting these screws in is that the assembly used to be welded. The welds failed by cracking along the HAZ. As welds create homogenous material from assemblies, the crack was able to run through the assembly along the interconnected HAZ lines and one whole side fell off. As a rule, unannealed welds are poor in high fatigue situations regardless of how many there may be in the world.

I think anything which is over-engineered and therefore sees little changes in stress will deal well with fatigue. Stiff welds surrounded by flexible sheet can make use of the sheet as a spring and last longer.

I still say that a work hardening material must become brittle if flexed in the yield zone. It may not happen quickly, but it must happen.