Preheat & PWHT Practice for Tube-Header Weld 1

[Guest102023]

I would like to know what is 'best practice' for repair of a single (or possibly several) tube-to-header welds.

I recently observed from a distance (I was only involved in the tube F/A) as a contractor proceeded to preheat only the end of the 12"Ø header where the repair was to be done, and with a propane torch. The repaired tube happened to be the last one, middle row; CSEF tube to conventional 2¼Cr-Mo header.

Firstly, I did not agree with oxyfuel preheating and temperature crayon monitoring - I thought now that we are in the 3rd millenium it would have been electric pads with TC control.

Second, I had doubts about thermal stresses resulting from PWHT only locally. I should think you would want to heat things up symmetrically.

Any experiences/practices anybody can share?

 

[metengr]

The above is not recommended for joining conventional Cr-Mo to CSEF steels because of poor temperature control.

For conventional Cr-Mo alloys joining a terminal tube to a header, local PWHT can be done with proper controls over the heating band and soaking band.

 

[stanweld]

brimstoner,
Agree with your thoughts and those of metengr. Do expect to make additional repairs to this repair! Oxy-fuel gas preheating has been shown to overheat the (assumed T91) tube. We have seen evidence whereby the T91 was heated above its lower critical temp. As a result, we prohibit its use for preheating T91/P91 weldments.

We would require electric resistance heating pads for preheating. While we would locally PWHT a single or a few repair welds near one another, we would control the temperature along the header to mitigate thermal stressing. A detailed PWHT procedure would be provided.

 

[Guest102023]

Thanks for both your comments confirming my thoughts.

In the situation where only one end of the header is repaired, to what temperature would you PWHT the header; i.e., the rest of the header?

 

[metengr]

If you are weld repairing only one end of a long header, I would use one or two additional resistance heater mats to keep the thermal gradient outside of the local heating band (repair region) to less then 250 deg F per foot. In other words, outside of the weld repair region the thermal gradient should be 250 deg F/foot in the header until 800 deg F is reached. At this point, metal temperatures of 800 deg F or less can be cooled in air for a header. Make sure the header can move axially and add temporary supports if needed.

 

[stanweld]

Take head to metengr regarding additional supports. On one such repair that mistakenly left off the added supports, it was not a pretty sight. Go to repair plan 2 and don't collect $200 if you pass Go.

 

[Guest102023]

The temporary supports would be required to reduce tensile loading on the tubes I assume. Possibly the CG-HAZ is of specific concern?

 

[stanweld]

Tensile load on the tubes was the problem. Note that on that repair, all tubes in the tube to header welds required PWHT. One side was not supported. Tilt!

 

[Guest102023]

BTW these tubes are T23, where SRC has been an ongoing issue.

 

[stanweld]

On T23 tubes often the locally applied preheat temp as you indicated is too low, resulting in HAZ hardness > 350 BHN. Also PWHT temp may be too low as well due to the "chimney" effect and the heat conductivity of the heavy wall header not subject to heating or properly insulated.

While I have not had the problem on T23 to P22 weld repairs, I had a very real problem when torch preheatinhg was done on some T22 to P22 tube to header welds. The hardness was near 350 BHN in the weld metal and P22 HAZ. After a leak was discovered during startup at a location not subject to the repair, the operator water quench cooled the system (thermal shock). From the failure investigation of the many cracked tube to header welds, it was believed that the actual PWHT hold temp was less than 1100F.

 

[Guest102023]

stanweld - What can be done to reduce the 'chimney' cooling effect, if the nearby system is closed up?

metengr - what form does the support take, simply some sturdy blocks between the header and the floor, or something more?

p.s., a tendency I notice with those who know 'a little' about welding metallurgy, but especially with heat treaters, is they feel PWHT of steel covers a multitude of sins and corrects all ills, especially incorrect preheat/interpass.

 

[stanweld]

The "chimney" effect is a dilemma. In general one cannot realistically cut interconnecting pipe and plug. Additional insulation applied along the tube and on the header is what we prescibe. We do consult heat treating experts who have shown their expertise in this arena to perform the PWHT. And we perform hardness checks to verify PWHT. When shown to be ineffective, an additional PWHT cycle(s) and added insulation is applied.

 

[Guest102023]

Additional insulation is also what I am hearing from my local heat treater. Not many internal access points in the business end of a HRSG!

You can only check the weld hardness unfortunately and not the HAZ, but that should still give some useful indication of PWHT effectiveness.

On the subject of repair, would you guys advise surface grinding of the weld before PWHT to eliminate any sharp stress raisers? I am guessing this is not usual practice (in this case it is 100% GTAW and the weld profiles generally look pretty good.

Thanks again.