thermal fatigue

[havok91]

I'm trying to understand the effects of thermal cycling on metal parts.
When a metal is heated it expands (I'll call it elongates) and contracts when cools. Simple stuff. However, what conditions must exist for a permenant set in elongation to occur? I assume that as the temp is less than the tempering temperature, the part will retern to its initial size. Is this true?
Your knowledge is appreciated.

Mike

 

[Neubaten]

A potential condition could be:

Combination of temperature, time and stress: Creep.

 

[EdStainless]

Or let's think about heating a part that isn't free to expand. You could exceed the compressive yield strength. Then when the part cooled it would contract, to a size smaller than it started out.
or a part might be free to expand, and then because of binding or interference it is restricted from contracting and gets stretched.
I have seen both of these. Neither one directly leads to thermal fatigue. TF usually results when parts are rapidly and non-uniformly heated and cooled.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[Neubaten]

By the way, I was referring to the conditions that could lead to permanent strain, not for thermal fatigue.

 

[knowlittle]

havok91,

Do you mean tempering like tempering after hardening? If that's what you mean, you are assuming that for T < tempering temperature, there will be no phase transformation, and therefore no plastic deformation. It is only partially true. As others said, uneven temperature and restricted thermal expansion result in plastic deformation. Get a good size metal plate, weld at all 4 edges, and raise/lower temperature of the plate (below tempering temperature). Soon you will have a twisted plate. I have seen steel plates cracking after a few years of repeated heating and cooling. I don't think those were true thermal fatigue failures all the way from beginning to end, but I am sure thermal fatigue played a part.

 

[Tmoose]

In cylinder heads, especially air-cooled, the temperature gradients are huge. As mentioned by others, local heating results in the hot section "upsetting" when it is constrained by a massive cool section. Then, upon cooling, large tensile stresses result. cracking after relatively few cycles is often considered low-cycle fatigue and requires yielding each cycle.

Legend has it that early and late BMW bike head made of "good" aluminum develop cracks in the combustion chamber, but resist spark plug and exhaust thread failure.

http://w6rec.com/duane/bmw/head/head1.JPG

http://www.aa1car.com/library/elements/cracked_cylinder_head.jpg

http://www.freewebs.com/bmw2002turbo/Cracked Head closeup.jpg

http://farm1.static.flickr.com/183/471411754_bad2f87ff2.jpg

 

[PEW]

That last photo is a real no-brainer.