Grade P11 Weld failure between ellipsoidal head and tubesheet 2

[qualitypro]

We encountered a weld failure between an ellipsoidal head and the tubesheet material in the field even before the equipment had been started up. The head in the new HX equipment was over 3 inch thick and had gone through all necessary preheat, NDE, PWHT, hydros etc in an overseas shop with no anomalies. What could be the potential causes for this failure? Appreciate if anybody can share any past experiences with HX weld failures.

 

[david339933]

No preheat maintenance? For this material and thickness?

 

[Christine74]

"Internal fillet weld is required according to table 4.2.6 detail 2"

I nominate r6155 to make the internal ellipsoidal head-to-tubesheet fillet weld.


-Christine

 

[r6155]

@ Christine74
Read my post again."Incorrect design. Internal fillet welding required according to table 4.2.6 detail 2"

Obviously, internal fillet welding is not possible, so this is an "incorrect design".

Try not to be sarcastic

Regards

 

[GD2]

OP,
Did you do a local hydrotest at your facility prior to commissioning and start up?

GDD
Canada

 

[Christine74]

The title of Table 4.2.6 specifically says "SOME Acceptable Weld Joints for Flat Heads..". The joint the OP is showing is perfectly acceptable per ASME and is routinely used on TEMA "N" and TEMA "C" tubesheet-to-cylinder attachment welds. No internal fillet weld is required.

But back to the OP: When RT is required for the joint you're describing, the usual practice is to use a forged tubesheet that includes a cylindrical stub with a projection at least a few inches in order to allow adequate space to conduct RT examination of the tubesheet-to-head weld.


-Christine

 

[r6155]

@ Christine74
a)The (OP) talks about ASME VIII Div 2.
TEMA standards are intended to apply only to exchangers designed to ASME VIII Div. 1 (see inquiries TEMA).

Table 4.2.6 Detail 1,2 and 3 are with inside fillet weld. Read more about CORNER weld.

b) “But back to the OP: When RT is required for the joint you're describing, the usual practice is to use a forged tubesheet that includes a cylindrical stub with a projection at least a few inches in order to allow adequate space to conduct RT examination of the tubesheet-to-head weld.”

Why not explain it with a reference sketch instead of words?

Can you explain how to make this 3” thick full penetration weld from the outside only?

Can you explain how to do RT if there is no access inside?

Regards

 

[Christine74]

These are all basic questions that have all been discussed at length in this forum already.

1) The channel head would already have nozzles installed in order for the process fluid to get into the tubes. If required handholes (or a manway) can be added. But no need for any of this if radiography isn't required.

2) See Figure UW-13(c) for a typical arrangement of tubesheet with a cylindrical stub I described above.

3) See Interpretation VIII-1-83-220 for one of several ways to make full penetration welds when welding from one side only.


-Christine

 

[r6155]

This is a common situation when the designer does not have manufacturing and inspection experience.
Example: see attached drawing, a 45° bevel is extremely exaggerated.
RT instead of UT?

@Christine 74
You insist on ASME VIII Div.1: incorrect
Butt weld 3” thickness P11 material, one side with preheating: impossible

Regards

 

[qualitypro]

@GD2, the vessel was hydrotested at the shop facility in the horizontal position per the ASME Data reports received. There was a shop shell side and tube side hydro done in conjunction without exceeding the maximum differential pressure as suggested by the OEM designer and signed off by the shop fabricators AI. The field piping to the HX nozzles were completely welded here and a field hydro in the vertical position, this being a vertical HX was done through the HX to verify integrity of the piping welds. Not sure if the horizontal vs vertical hydro could altogether be different at this bottom dish weld joint and if the max differential pressures between the tube and shell side were maintained.

 

[qualitypro]

All, in relation to this repair to be undertaken by a fabricator, will a R stamp holder be enough or do they also need to have a U2 stamp?

 

[Christine74]

An "R" stamp is all that would be required to make the required repair.


-Christine

 

[DekDee]

"the crack starts in the weld centerline from its initiation point and went circumferentially for abt 5 feet along the weld before it propagated diagonally into the dish head for about 3 feet."

8 foot long (2.4 mtrs) crack - and you are going to try and repair ?
Good luck.

 

[r6155]

I insist with my post 10 JAN 24 11:45
“What does the manufacturer say?”

If you don't know how the failure occurred, nothing can be repaired.
DON'T YOU HAVE THE AI REPORT YET?

I (as a user representative) had a serious problem with the AI. No inspection is carried out during manufacturing, detailed drawings are not verified as a minimum scope of work. Manufacturing was stopped until the AI and the user were brought together in the workshop. Conclusion: waste time and money with AI.

It seems like a similar situation to me now

Regards

 

[qualitypro]

@R6155, this will be reworked to NBIC R stamp. The bottom dish and the groove weld will be replaced in its entirety. We are in the process of getting a new dished end from the OEM but the important aspect is with limited access from the inside and with a single bevel weld at this 3" joint configuration how can we ensure a good root weld can be achieved. WE arec reaching out to the OEM to understand if the joint design can be changed so we dont end up with a similar issue again after the repair work.

 

[r6155]

@ qualitypro
1)How do you plan to repair the damaged area on the tubesheet after removing the weld with the semi-elliptical head?

2)How do you plan to repair if there is a leak at the tube to tubesheet joint?

3)Also think in Naroow gap welding technique.

 

[GD2]

[highlight #EDD400][highlight #BABDB6]@GD2, the vessel was hydrotested at the shop facility in the horizontal position per the ASME Data reports received. There was a shop shell side and tube side hydro done in conjunction without exceeding the maximum differential pressure as suggested by the OEM designer and signed off by the shop fabricators AI. The field piping to the HX nozzles were completely welded here and a field hydro in the vertical position, this being a vertical HX was done through the HX to verify integrity of the piping welds. Not sure if the horizontal vs vertical hydro could altogether be different at this bottom dish weld joint and if the max differential pressures between the tube and shell side were maintained.[/highlight]

Thanks OP for responding.
So two hydrotests were done on this HX. I will be more interested on the hydrotest temperatures. Can you tell us the test temperatures?
The horizontal and vertical hydrotest shouldn't make difference as most vertical vessels are shop tested horizontal.
I am sure the HX have been designed for vertical installation.

GDD
Canada

 

[qualitypro]

@GD2, the test temperature was to be a minimum of 60 F and this was maintained both at shop and field tests. This is a vertical HX.

 

[qualitypro]

The 3" thick dish material, ASTM A387 Gr.11 Cl.2 MTR calls for normalizing with a soak time of minimum 71 minutes. The dished end was hotformed and then normalized for 189 minutes per the shop HT charts shared. Any insights into if this is acceptable.

 

[r6155]

1) For heat Treatment the complete data is: heating rate, holding time and cooling rate
2) Test temperature is the "metal temperature".

Regards

 

[FMJalink]

Hi qualitypro,
Looking at the challenge you got, I see 3 major points of worry and would better be investigated prior to a re-attempt. First: the collar on the head is, looking to the wall thickness of the head, quite small. This leads to high local stresses on the inside due to head bending with pressure on the inside. Second: the 45 degree opening with 3 inch wall thickness will lead to a lot of remaing stress. As you will know, each new layer (from inside to outside) of weld material shrinks relatively to the previous layer. Somewhere heat treatment might be required. Maybe better to use another weld shape. Third: I do not see specification of the vertical gap on the root of the weld. In order to have a full penetration weld, the welding material needs to get into the far end òf the weld opening. I doubt whether this wedge shape would even allow this to happen.

If I understood it correctly, an attempt will be undertaken to replace the head. That will require also a lot of preparation and thinking with this wall thickness and material type. I have seen not long ago how wrongly opening thick welded parts, made scrap of the heat exchanger and that heat exchanger was not made in this more difficult CrMo.

Succes.