Stack weld cracking

[cespi82]

Hello everyone,

I have this issue. We are under the commissioning process of one Gas Dehydration Unit. 6 months ago we started to conduct some tests in the unit as a part of the commissioning process and after two days testing the unit at the operating conditions we realized that we had a design error in one equipment. We shut down the whole unit, flushed out and cleaned all the equipment and pipes while this defective equipment was being repaired. The only item we did not take care of during this idle time was the fire box and the stack of the Glycol Reboiler.

Once we got the defective equipment fixed and installed back in the unit, we started again with the test and after about two hours of the burners start up we experienced a crack in one of the welds. Please the picture attached for your better understanding of the situation.

The Fuel Gas H2S content is 400 ppm average. The materials of the fire box and stack is CS.

My main suspicion is that the combustion gases (SOx ) in the fire box and stack condensated to form Sulfuric acid and sulfur acidic compounds, and caused some corrosion damages internally, but I wouldn't expect cracking. The other option could be a defective fabrication/welding.

I would appreciate your thoughts on what could happened here...

Kind regards,

Cesar

 

[weldstan]

When improperly made/sequenced, the fillet weld in question may well have been cracked or partially cracked during manufacture, transport or handling. I do not suspect corrosion as the culprit. With the observed corrosion products along the crack and on the surface below the crack, your suposition that condensation occurred appears quite valid.

 

[metengr]

The crack clearly follows the toe of the weld full length along the plate and it appears to have displaced the lower plate inward after final fracture. The displacement is confirmed because a portion of the weld fracture surface is exposed. The location and appearance of the crack suggest the plate was under significant stress. I would not necessarily call this defective welding. I would agree corrosion did not play a part in the failure.

 

[cespi82]

Thank you very much for answering Weldstan and Metengr...

Ok, we all agree that corrosion is unlikely to be the cause of this failure.
The point of considering a failure during the welding, from my standpoint, is because this chamber is not operating under high pressure, it is actually open to the atmosphere, so what such big stresses can be imposed to this weld joint to make it crack?

Metengr, may be your are seeing something I don't see about where those stresses might come from. Might be related to the idle time? Might be flame impingement? if the plate moved inward, that could mean there was a negative pressure inside this chamber?

On the other side of the coin, any suggestion/consideration to repair this? As of now, i'm only thinking in wash the weld out and make the fillet welding again after performing some Penetrant Testing on the surfaces. I will do PT in the rest of the welds around, I hope not to find other cracks.

Kind regards and thanks again!

 

[weldstan]

The stress is induced during fit-up and welding. Welding directly into sharp corners can induce stress high enough to cause cracking. While the weld's profile appears acceptable, if fitting, tacking, and sequencing are not adequately addressed, cracking may occur. I suggect that you determine just how the weld was made to determine best repair methodology.

 

[boo1]

We had cases in turbine exaust collectors cracking. We tried changing our welding process, even annealed the assy and reheat treated it. The cracking was was not effected.

Our study found the thermal changes created stresses, and cracked the collector. We changed our design to allow for the growth.

 

[weldstan]

Design may be the culprit or a highly contributing factor. How thick is the flange relative to the box structure? Great thickness differences can lead to large thermal gradients resulting in high stresses during startup and shutdowns.

 

[dik]

Triaxial stresses in the corner could be the problem; I'm not sure of the procedure to minimise/eliminate these. It could generate a brittle crack that initiates the balance of the crack observed. What are the materials and the welding process?

Dik