A question about why welds fail on a manifold

Trboatworks · Aug 19, 2018

A question about why welds fail on a exhaust.

Ok- a pair of exhausts ran through my buddies marine fabrication shop for repairs.
I didn't get a photo of the damage but the weld cracked right down the middle on that shoulder where the 3" was stepped up to 4".
Both exhausts had the same crack

We were talking about what caused the crack and he was thinking that the heat got to the weld and was the cause.
I am thinking more a vibration thing.

So... I tend to overthink these things at times
My thinking has a vibration frequency dropping as one moves away from the turbo where this bolts up to the lowest frequency out at end where soft hose is hooked up.
I am seeing a vibration node caused by that step up from 3" to 4" on that shoulder which has a sudden shift in frequency and loads up the metal causing fatigue failure.
So the frequency is dropping as we move away from the flange but that drop is not linear.
The crack is found where the greatest change in frequency occurs at that shoulder.

I would love to see a heat map on a part like this to show stress risers but am betting I am right- that shoulder is one and the crack has more to do with that vibration being concentrated there than heat.
I guess my next question is why it wouldn't crack right at the flange where the stress is greatest and not the shoulder..
This sort of question makes me wish I had some engineering theory under my belt..

Thoughts?

I drew a line where the crack was on the second photo:

.

metalpyro · Aug 19, 2018

I would inquire if the engine is mounted solid or if there are any rubber mounts to isolate it from the hull. Same thing with the mount on the soft hose adapter. If it is solid, My gut says maybe a thermal expansion issue, with both ends being restrained. If rubber mounted, I might be thinking it ends up in a binding problem when the engine flexes it's mounts.

Overthinking? Nah.
It's what we do.

Trboatworks · Aug 19, 2018

metalpyro said:
I would inquire if the engine is mounted solid or if there are any rubber mounts to isolate it from the hull. Same thing with the mount on the soft hose adapter. If it is solid, My gut says maybe a thermal expansion issue, with both ends being restrained. If rubber mounted, I might be thinking it ends up in a binding problem when the engine flexes it's mounts.

Overthinking? Nah.
It's what we do.

Thanks for that- I will check on mountings.

Re-overthinking.

In my business a great deal of what one might call success lies in having the experience to correctly call what has caused failures in gear.
It does little good in some cases to carry on and buy a bit more time in a quick repair if the part is headed towards the same failure.
Things being what they are the builder is facing the same sort of challenge as failures are common on gear and structures and a reputation is made on eliminating those failures by correct assessments of causes.

In Kinney's Elements of Yacht Design there is a nice anecdote where a worker calls out a failure mode correctly as the naval architect mutters under his breath; "don't give a opinion unless you are paid for it"...

For me this sort of problem is a daily occurrence- there is a plate power rudder in the yard just now suffering from vibration with no play in bearings- prop pounding, fairing issue? While another is tearing out its interior under way due to too rigid of panels on a flexing cabin.

All part of the game but yeah- I have some ideals which are not worth much- the frequency node as above might be one of them.

adama · Aug 19, 2018

Vibration - constant slight flexing, prop non-alignments etc its a common thing in the marine world, we even do them for the local life boats, some boats are noticeably worse than others. Real fix is ditch the stainless and go steel, yeah they rust through, but actually tend to last longer especially if you use heavy pipe! Think thats also the isue, most the stainless ones try and use 16G or thinner tube like on a car, they realy want more mass - strength to em.

Lots of smaller boats leave a lot to be desired in the prop - engine alignment too, most boats are pretty floppy, waves chop and slight temp change all takes things into and out of alignment.

Trboatworks · Aug 19, 2018

Yep it's stainless and light gauge as you say.

hobbyman · Aug 19, 2018

Stainless has a low fatigue life. Is that a jacketed exhaust? Stainless? I am wondering if it had been passivated after welding and if the inside had been backed by inert gas.

johansen · Aug 19, 2018

If the steel has a constant cross section from the 3" through the coupling to the 4" section and if the welds are the same, it would have failed at the 3" weld, not the 4" and it would have broken off at the flange, not the coupling. (unless the vibration that broke it is a much higher frequency)

Due to the relatively low stiffness 3" pipe compared to the 4", all the bending is going to be in the 3" section unless the 3 inch section is 3.1 times the wall thickness of the 4". (which it could be) and for the same strength, the 3 inch pipe needs to be 2.3 times as thick.

its also likely the coupling is thicker than both pipe sections.

so my guess is bad weld on the 4" section.

Trboatworks · Aug 19, 2018

Thanks- that makes a lot of sense- the smaller pipe section is less stiff and will flex more so reach fatigue faster.

The part I am trying to wrap my head around is how the material responds to when that flex is constrained as it meets the larger diameter/more rigid section.

It seems that junction represents less flex but more yield therefore faster fatigue???

Just mulling this over.

My buddy told me the cracks were right down center of weld - same place both exhausts.

Edit- rereading what you said.
Yes- should crack at the 3” weld ..

thermite · Aug 19, 2018

Trboatworks said:
Ok- a pair of exhausts ran through my buddies marine fabrication shop for repairs.
I didn't get a photo of the damage but the weld cracked right down the middle on that shoulder where the 3" was stepped up to 4".
Both exhausts had the same crack

We were talking about what caused the crack and he was thinking that the heat got to the weld and was the cause.
I am thinking more a vibration thing.

So... I tend to overthink these things at times
My thinking has a vibration frequency dropping as one moves away from the turbo where this bolts up to the lowest frequency out at end where soft hose is hooked up.
I am seeing a vibration node caused by that step up from 3" to 4" on that shoulder which has a sudden shift in frequency and loads up the metal causing fatigue failure.
So the frequency is dropping as we move away from the flange but that drop is not linear.
The crack is found where the greatest change in frequency occurs at that shoulder.

I would love to see a heat map on a part like this to show stress risers but am betting I am right- that shoulder is one and the crack has more to do with that vibration being concentrated there than heat.
I guess my next question is why it wouldn't crack right at the flange where the stress is greatest and not the shoulder..
This sort of question makes me wish I had some engineering theory under my belt..

Thoughts?

I drew a line where the crack was on the second photo:

View attachment 236155. View attachment 236157

That ain't a "manifold' in the picture. "Connector", "runner", "riser" wotever - it has but two ends, not a branching-out to "many" as-in manyfold.

Reason it matters is that a true "manifold" the weld failure mode is heavily tlted to uneven thermal stress, and no place for it to relieve itself by moving.

On a simple item, here, thermal stress is able to be released by movement. Vibration stress is possible. RAPID vs slow cool-down (highly UNeven stress) due to heat trapped one place, another place not-so-much is another suspect.

No expert on marine engines. Got more experience than I WISH I had on custom-welded tube headers, land-based.

FWIW- the weld you pointed to was a stress-riser from Big Bang.

Try to have a weld do ONE job, not two wherever idle to full-gallop entails an extreme temperature swing.

Swedge a curve into the end to bring it down gently ELSE flare the smaller tube UP gently, rather than weld a thickened joint right into a hot transition zone. Differential stresses of cool-down the killer, here.

Either change the profile or make it a slip-fit and Nickel braze that. Or all of the above.

kustomizingkid · Aug 19, 2018

That pipe wants to flex and it can't...

https://vibrantperformance.com/cata..._1254&osCsid=d85aaf7f53e02453e6d13b907c8e8ca9

Trboatworks · Aug 19, 2018

thermite said:
Differential stresses of cool-down the killer, here.

Thanks.

Bill- thread title was misstated- exhaust part...

thermite · Aug 19, 2018

Trboatworks said:
Thanks.

Bill- thread title was misstated- exhaust part...

Thanks, I grokked that.

The difference in where one looks "first" seemed worth mentioning.

If you 'section' that one you have with a saw, you'll see a markedly thicker ring formation where the joint and weld are, vs the thin sheet's thickness. And then- one side of it is at sharp angle, so it isn't acting like a thinwall cylinder. Approaches a radial loading, rather. Now they are at war with each other.

DDoug · Aug 19, 2018

thermite said:
If you 'section' that one you have with a saw, you'll see a markedly thicker ring formation where the joint and weld are, vs the thin sheet's thickness. And then- one side of it is at sharp angle, so it isn't acting like a thinwall cylinder. Approaches a radial loading, rather. Now they are at war with each other.

Yup, sharp change in material thickness.
un-even heating, vibrations, or weld problems (note pix looks like crack is
just off the weld, but still in the HAZ)

cyanidekid · Aug 19, 2018

the faulty assumption you may be making here is that all is equal but for the stiffness and cross-sectional area, and how the vibration is channeled through it.

in fact, because the part is non-homogenious, and incorporates welds, harmonics and vibration propagation are potentially MUCH less important than crystal structure/grain structure, including in the HAZ (as digger points out). it is the weld it's self and how it differs from the metal adjacent that is the weak point.

you have no good info on the parameters/quality of the original welds, even what filler metal was used, so you can't know how that affects the failure mode.

there are just too many variables in all the above to make any good analysis of it.

as to the fatigue resistance of stainless, austenitic alloys (300 series, particularly low carbon varieties) would have considerably greater resistance to fatigue than carbon steels or 400 series stainless alloys (if properly welded).

yea, carbon steel would be more resistant to cracking in a normal exhaust application, because it would rust through first.

Illinoyance · Aug 19, 2018

Cut a coupon at right angles to the weld. Polish and etch with acid. That will yield a lot of information: weld penetration, location of the crack with respect weld or HAZ. Until you see the cross section of the weld you can't rule out defective welds or metallurgical problems.

thermite · Aug 20, 2018

Illinoyance said:
Cut a coupon at right angles to the weld. Polish and etch with acid. That will yield a lot of information: weld penetration, location of the crack with respect weld or HAZ. Until you see the cross section of the weld you can't rule out defective welds or metallurgical problems.

Yes. You can. This profile ain't gonna work. Stresses are just wrong.

Get it "right" you don't even need welds. Slip joints, clamps, burnt-off slather of linseed oil & the carbon and carbonized varnish that produces wedges in the gap and seals 'em right up. Rust can come in from the back side, lock 'em so tight you don't take those apart, you cut them out and start fresh.

You only have a few inches of water, back pressure, for a joint to stand before it is easier for the gases go on down the line and OUT, after all.

No relationship to a water line or hydraulic line's pressure. Closer to DWV sanitary sewer, just HOTTER and with combustion hammer-pulses & resonances.

johansen · Aug 20, 2018

thermite said:
Yes. You can. This profile ain't gonna work. Stresses are just wrong.

Get it "right" you don't even need welds. Slip joints, clamps, burnt-off slather of linseed oil & the carbon and carbonized varnish that produces wedges in the gap and seals 'em right up. Rust can come in from the back side, lock 'em so tight you don't take those apart, you cut them out and start fresh.

tempted to put thermite on the ignore list but the humor is almost worth it, especially after a few :drink:

(s)

DDoug · Aug 20, 2018

johansen said:
tempted to put thermite on the ignore list but the humor is almost worth it, especially after a few (s)

That was quite ignorant of you.

Look at a piping fitting catalog, note how some remove the weld from
the highly stressed area's, much like He is advocating.

thermite · Aug 20, 2018

digger doug said:
That was quite ignorant of you.

Look at a piping fitting catalog, note how some remove the weld from
the highly stressed area's, much like He is advocating.

LOL! Never mind the metallurgy texts on my shelf. It was actually a Bulgarian born local artist, Mannasas, VA who crafted the tubes for my uber-Jeep "sleeper" that taught me the most.

I "sneak" it over to his shop wee-hours, nothing aft of the tube headers Advance had built to my full-scale model, windows wide-open to not succumb to the fumes.

George puts it up on the lift next morning, walks under it EYEBALLS it, starts up his powered tubing mangling machinery, lays the whole system in by eye, never even picked up a tape-measure but exactly TWICE. Its was what he DID, all-day, every day.

An "exhaust guy".

Michelangelo wudda struggled to match his calibrated eye, lost his very ass on the speed of it all, too.

Johansen's problem?

Long on maths.

Short on understanding.

Just made that abundantly clear - again - on projectiles, gunsmithing forum.

And THAT one was seriously funny!

johansen · Aug 20, 2018

Lol. Last time i tried i put 9 out of 10 into a dinner plate at 500 yards.

A question about why welds fail on a manifold

Diamond

Aluminum

Diamond

Diamond

Diamond

Hot Rolled

Stainless

Diamond

Diamond

Titanium

Diamond

Diamond

Diamond

Titanium

Stainless

Diamond

Stainless

Diamond

Diamond

Stainless

Similar threads