Cracked weld puzzler

[Watermelon]

Hello.

I have a 20" pipe, 30' long, 1" wall, for use in some industrial process equipment. While the pipe is rotating, heavy process material is dropped on it, normal to the major axis. Loads are substantial and shock/impact applied. The rotating pipe is subject to significant bending type loads. It is a hostile environment.

On the pipe there are numerous circumferential collars (donuts)which support teeth. The collars are 2.5" thick, stand 2" high from the pipe and occur every 9" along the length of the pipe. Each collar is welded to the pipe by 1/4" fillet welds on each side. Each fillet is 6" long. The welds occur generally at 12, 3, 6 and 9 o'clock.

Numerous welds have cracked. Most of the cracking occurs at midspan (center) of the pipe. Some welds have cracked down the center, others have cracked at the toe of the fillet on the pipe. The crack follows the pipe circumferentially and extends into the pipe on the pipe cross-section. The metals have been tested and are certified as weldable grade materials.

I'm sure the problem is related to bending and fatigue, but also to welding procedures.

Questions: What is the most likely source of the problem?

What kind of welding procedures can be used to help prevent future problems?

Muchas gracias for your help.

 

[diamondjim]

If you are getting most of cracks near the center
of the pipe it may be the result of torsional
stresses. What is driving the pipe? Does the
pipe turn in one direction only? Is it rotating
like a sewer pipe in that the primary axis is
horizontal? Interesting problem! What kind of
speeds? A 1/4 weld is not much for heavy impact
loads.

 

[Watermelon]

Hi Diamondjim:

I'm not sure about the mechanical items (I didn't design it). I believe it's chain and sprocket driven, perhaps 30 rpm, one direction only. It rotates horizontally about its major axis, like the rollers in a plotter. (I was going to say typewriter, but that dates me.) I'd be surprised if there was no torsion involved, because of jamming from time to time.

I think your last line is a key statement. A 1/4" fillet does not seem appropriate given the thicknesses of the materials involved.

Thanks very much,

Watermelon

 

[pipewelder1999]

The intermittent fillet welds would be more subject to failure from impact/fatigue loads than a continuous fillet weld. A continuous weld would be more resistant to this type of failure based on my knowledge. This is based on information gained through AWS Welding Handbook Volume 1 and Lincoln Electrics Design Of Welded Structures.
Gerald Austin
Iuka, Mississippi

http://www.weldinginspectionsvcs.com

 

[arto]

Is temperature involved?
One time we had a stainless steelpressure vessel with carbon steel rings that had humongous stresses from differential thermal expansion [Stress corrosion cracking]

The old Roark book has a case for a ring on a shell under pressure. [3rd ed, Table XIII, case 9] solve for equivalent pressure to get same radial deflection as thermal gives you,then go on from there.

 

[boo1]

Without the loading or strain guage data, type of steel / heat treat, carbon content, and weld rod used I offer the following ideas:

The added plate collars increase the section modules in the localized area.
Bending loads are being transmitted through collars.
Disconuities from collar to pipe, causes stress riser.
In addition to shock, fatigue should be considered.
Sresses not in complete reversal.
Natural frequency of member not unknown.

Weld below mininum fillet size, for 2 1/2" plate use min 1/2" weld (AWS)to prevent too fast a cooling rate through thick plate or preheat and use 3/8" continous weld.

 

[GRoberts]

Do the welds recieve a stress relief PWHT? With all the cyclical loading that seems to be taking place, it is common practice to make SR PWHT mandatory so that the HAZ of the weld, which is harder than the base metal is tempered to a lower hardness and reduced the stress riser caused by the hard HAZ. This may be less of a problem than the weld size and intermittancy mentioned in other earlier posts.

 

[Watermelon]

Thanks to all for the responses.

I'm on my own now. If and when time permits, I'll post a follow-up memo.

Great web site.

 

[boo1]

Fracture can be classified into two general categories, ductile and brittle.
(1) Ductile fracture occurs by deformation of the crystals and slip relative to each other. (failure at 45 degrees plane)
(2) Brittle fracture occurs by cleavage across individual crystals.
There are four factors that should be reviewed when analyzing a fractured surface. They are growth marking, fracture mode, fracture surface texture and appearance, and amount of yielding or plastic deformation at the fracture surface.

Welds crack for many reasons, including the following:
(1) Insufficient weld metal cross section to sustain the loads involved.
(2) Insufficient ductility of weld metal to yield under stresses involved.
(3) Under-bead cracking due to hydrogen pickup in a hardenable type of base material.
(4) hot cracking or cold cracking
(5) stiffness or rigidity of the weldment itself

The total joint strength must still be great enough to meet design requirements. The residual stesses in the heat-affected zone effect the joint strength. Stress relieving of large structures in the field are difficult.

In this case I suspect the stresses applied to the joint exceed the yield strength.

 

[sbc]

I agree with pre-heating and full welding as posted
above, but it sounds as if these teeth may need to
be replaced periodicly?
One solution may be to roll mild steel to form a
collar of sufficient width (say 7&quot, weld it to
the pipe, then weld your 2.5" wide collars to it.
You may need to machine the I.D.s to accomodate.
This will add rigidity to the pipe and provide
some isolation of the shock loads.

 

[todousa]

If you are welding with a low hydrogen rod or wire you may want to check your welding procedure which should include pre-heating requirements. Also, from what you've described you may want to re-visit the weldment design. If you can increase the weld size to 5/16 or 3/8 using a more impact resistant weld metal it may make the cracking problem go away. Another way to increase your affective root size is to make the weldment a partial pen weld instead of a fillet weld. In any case I suggest you visit the "Lincoln" welding web site for more information on weldment materials.