Thickness of gasket influence on preload?

[danohl]

Does the thickness of the gasket influence the preload when temperature is changed?

I have a bolt connection where everything is steel except the gasket (it is actually a composite part). Everything is preloaded so that the bolt is under tension. When temperature is changed the gasket will expand differently than the steel due to the different thermal expansion coefficient and stiffness. Since the therm. coeff. (alpha) is length/length/temp; does it matter if the gasket is 1mm or 0.1mm? What am I missing?

 

[MikeHalloran]

The gasket also shrinks with falling temperature, which is probably more important to bolt preload.

And the second /length/ in the units refers to the gasket thickness, so a thinner gasket will affect bolt preload less. There may be other reasons for preferring a thick gasket, but bolt preload is not one of them.

One decent example of a thin gasket used to maintain bolt preload is in manual transmission gearboxes, e.g. between case and tailshaft housing, or between case and input shaft housing; traditionally made of specialized thin paper.



Mike Halloran
Pembroke Pines, FL, USA

 

[Tmoose]

CTE predicts how much thicker the gasket would like to be when hot, but the modulus of elasticity determines how hard it will push trying to become that thickness, and the various strengths will determine if it really pushes that hard, or just creeps or "upsets" internally, and will be thinner upon cooling.

I'd be frightened of preload loss in a structural joint that includes a gasket, or whose mating faces are even painted.

 

[MintJulep]

The key is to look at the amount of bolt stretch due to preload, and compare that to changes bolt length and thickness of clamped parts due to all factors.

If the changes exceed the original stretch then all preload is gone.

 

[rb1957]

CTE is like strain, it's a factor to be applied to a length (in your case the thickness) ... a thicker piece will have more expansion/contraction. i'm guessing that the gasket is a pretty small thickness (<0.04" ?) compared to the bolted flanges (> 0.125" ?), i'd've thought this was a pretty small effect, lost in the noise ??

 

[danohl]

Thanks for the replies,
I can't really let go of the problem. Your arguments feels logic but I just can't prove it... What I'm looking at is a bolt connection of composite panels so the thickness of the "gasket" is quite high.

If I assume for one material: F = -k*dx (a linear spring).

Let's say dx = a*dT*t, which is the distance, delta, the material compresses or extends (and where a*dT is the strain [m/m]). And finally t is the thickness [m].

I then say the stiffness k = E*A/t for a typical bar (t being the length of the bar), E the stiffness [N/m^2] and A the cross section [m^2].

Then I get F = (E*A/t)*a*dT*t = E*A*a*dT

Which step am I missing here?

 

[rb1957]

you're quite right, the gasket is compressed by the bolt load, just like the flanges. and so will also be affected by temperature. it's just that it is (should be ?) a very small factor in the overall loading of a bolted joint. CoF of the bolt, and means of tightening/preload control, have a much bigger impact on the range of the preload achieved.

 

[MikeHalloran]

Ah. I thought you were talking about a gasket/flange assembly like, say, an engine's cylinder head.

If you've got a soft gasket, two composite flanges, and two thin steel flange reinforcing compressed, that's a whole nother thing.
... like valve cover gaskets, which are basically impossible.


Mike Halloran
Pembroke Pines, FL, USA

 

[desertfox]

Hi danohl

This site might help:-

http://school.mech.uwa.edu.au/~dwright/DANotes/threads/joints/joints.html#top

desertfox

 

[danohl]

Yes, basically that's what I have. Actually I have no gaskets, only composite panels bolted to a steel frame (washer-composite-steel frame-washer). But since the stiffness is much less than for steel I guess it can be compared to a gasket.

The idea is that for daily fatigue loading all in-plane loads are transferred by friction, hence I need some preload.

rb1957; You are probably right, the values will be very small. I just need to document that they are

Thanks for all inputs

 

[MikeHalloran]

If you need other than trivial preload in the clamping bolts, then consider bonding thin steel sleeves into the mating holes in the composite, so that the composite is not part of the stackup resisting the bolt preload.

Mike Halloran
Pembroke Pines, FL, USA

 

[rb1957]

wouldn't that pretty much defeat the purpose of the gasket ?

 

[MikeHalloran]

Depending on tolerances, it should result in controlled but limited compression of the gasket. ... but there is no gasket, according to the latest bread crumb.


Mike Halloran
Pembroke Pines, FL, USA

 

[MintJulep]

The relevant terms to search for are "embeddment" and "relaxation".

In broad strokes, soft materials clamped in a bolted joint are generally bad. Then tend to "ooze", getting thinner over time.

If the get thinner enough much or even all of the stretch of the bolt that was introduced by the initial preload will be lost and the joint will fail.

Long stretchy bolts are good. Short stiff bolts are bad.

Work out the bolt stretch from whatever preload you feel you need.

Figure out how much loss of thickness of the panels is possible over time. This may be a difficult thing to do.

If the bolt stretch is very much larger than the potential amount of relaxation then all is good. If not you've got a problem.