Leaks in seal welds and socket welds

[andyenergy]

I'd like some views from some of the experts on this forum if possible.

Bearing in mind I'm just about to go on holiday for a couple of weeks, I won't be able to respond to any queries you may have but I'd like to gather opinion in the first instance.

I have been seeing a possible increase in leaks developing in seal welds over the past couple of years. I have no statistics or data to back this up but simply observations of the number of times that I've been called on this topic.

These can happen in a variety of materials, but typically are in carbon steels or stainless and are in such geometries as tube to tube plate seal welds, plugs that are inserted in tubes to blank heat exchanger paths. These tube blanks usually consist of a solid prepared plug insterted in the bore of a tube that is typically of the order of 19 mm OD with a wall thickness of 3 mm maximum.

These welds may be one or two pass TIG (GTAW) welds and the leak usuaully manifests itself as a pinhole leak. The geometery and access is usually such that investigation of the root cause is difficult if not impossible although some rectification has revealed sub surface pores that are bigger than the leak site or exit hole.

In some cases, seal welds have been successfully surface inspected with recordable defects although again, access limitations often dictate that the surfaces are as-welded which has an obvious limitation in terms of the size of defects that can be detetcted. In addition, low pressure testing confirms the freedom from through weld defects but leaks develop either in service or as result of high pressure boiler filling (at a much higher presure than the original leak detection phase).

I belive that there are at least a couple of contributory factors - cleanliness of the joint and access. Both of these may produce conditions whereby a seal weld may look OK on the surface and resist a low pressure leak test but may contain sub surface defects that have a small remaining ligament which would fail a high pressure test.

In addition, thermal or pressure cycling can cause these ligaments to fail in time. Our applications are generally limited to very few cycles of either pressure or temperature. Temperature ranges are quite low and are associated with feed or economiser temperatures (100 - 200C. Other failures I have seen may have been in low alloy materials where creep may have also contributed but the root cause of sub surface defects is likely to be the same.

I would be very intertested in any feedback from others who may have encountered similar issues and any published information or operational experience on any testing that may have taken place. Feedback on my thoughts of the likely mechanisms would be gratefully received.

As I will not be around for a couple of weeks, I look forward to any responses when I return.

Thanks for your time.

Andy

 

[metengr]

Andy;
Enjoy your holiday. I have not really seen this problem of increased leaks on our boiler components at the welded locations(appurtenances or drain lines, etc) you mentioned.

We typically go through periods where after outages you have workmanship issues that appear. Once these are taken care of, things settle out.

I would say that if I had to rank the number one and number two contributing factors to seal weld and socket weld leaks, it has to be operational (cycling) followed by fit-up for socket welds. I have seen various boilermakers (welders) not understand that for socket welds, the gap required for stand-off distance is absolutely critical. I would bet that pin hole leaks you see are caused by root cracking (ID to OD surface) because the pipe nipple was bottomed out in the joint when the weld was deposited.

Seal weld leaks are probably more indicative of thermal/mechanical stresses related to cycling conditions in service. Not much you can do here other than blend grind weld toes to remove undercut and pay attention to workmanship detail.

Seal weld leaks in high temperature components could be indicative of a more serious problem related to creep deformation in service. This is especially true for gamma plug (RT) seal welds.

 

[metengr]

I did forget to mention that we have seen rashes of seal weld leaks on our lower drum hand hole caps. These caps are normally rolled into the header or drum followed by a seal weld. I have performed failure analysis years ago when we observed a series of hand hole cap leaks in service.

As it turned out, the leaks were cause by corrosion fatigue damage. The high stress regions of the caps during poor boiler layups resulted in conditions that could initiate and propagate corrosion fatigue cracks that eventually propagated through the seal weld or even through the cap itself resulting in a leak.

These leaks occurred during a period of time in the late 1980's where we had low capacity factors on the boiler and it cycled every other day or week. After we replaced the hand hole caps and started to use nitrogen for lay-ups, and our boiler capacity factor increased to 85-90%, this problem went away.

 

[deanc]

metengr has pointed out one problem. This can cause you problems in seal welded threaded items as well.

Few thoughts.

If the leak is on the stop the welder may not be ending the weld properly. One way is to fill the crater and stop adding filler,then draw the arc back over the weld.

Foot pedal helps to taper down the current.

If you are plugging something that has been in service you may need to cut away part of the tube so you can get to clean metal.

You could also have gases expanding due to the weld heat and blowing out your molten metal. Preheat will help.

Preheat will also help get rid of the crude during cleaning.

 

[stanweld]

We have found that lack of welder training in proper tie-in technique has been the principal cause of seal weld/socket weld leaks. When proper techniques are used, the incidence of failure during initial hydrotest are nil and greatly diminished during cyclic service.

With regard to tube-to-tubesheet welds, I recommend qualification of the welders on mock tube-to-tubesheet joints.

 

[0707]

In socket welds, to avoid cracks, is very important to guarantee that you will have a gap for expansion between the socket and the pipe. Normally this gap shall be 1 to 1.5 mm.

luis marques

 

[grampi1]

Never allow single pass socket welds. Its bad practice and the classic failure mode is a pinhole leak.

 

[andyenergy]

I'm back, refreshed but wishing I was still on that boat in the Indian Ocean....

Thanks for all your contibutions.

Many of the observations are similar to thoughts that I'd had myself.

Metengr -


Many of our problems also relate to workmanship issues and the difficulty in carrying out an adequate surface inspection and with no possibility of a sensible volumetric inspection.

The old story of the gap in sockets.... I'm aware that this is an issue and it raises the question of whether to radiopgraph or not. If the gap is still there, then the socket was properly set up. If it isn't then it may have been properly set up or not. Difficult to tell. Interestingly, a relatively new metallic pipework standard in Europe requires percentage radiography for sockets when operating above a certain temperature to demonstrate the gap is still there, presumably to ensure that an adequate gap is there to allow for thremal expansion.

Seal weld leaks when cracked have tended to be centreline and not associated with weld toes - this does indicated cyclic stressing from the root.

The issue of corrosion fatigue is one that I hadn't thought of - I'm not sure we have the right conditions for this but it's certainly worth bearing in mind.


Deanc - the welder workmanship certainly may be a significant issue. I am not convinced that many welders gives as much thought to apparently simple welds as they might. As for access and cutting tubes back, this might make the welds more difficult. Crud in the interface is certainly an issue and likely to cause problems. When i am around and these things are being welded, I'm often surpirised and dismayed by the lack of attention to removal of moisture and corrosion products.

Stanweld - similar comments to above and I also endorse the use of mock ups in difficult access situations for these types of welds.

0707 - covered by my reply to metengr above

Grampi - All seal or sockets are now specified as two pass, with the second pass intended to isolate the first pass and any problems that may be lurking sub surface.

In conclusion, I think that even if a weld looks OK on the surface it may well be harbouring a sub surface defect due to porosity resulting from moisture, expanding gas, rubbish in the interface etc. The answer is to ensure cleanliness, dryness and to deposit two passes always.

Thanks very much for your contributions, they are all gratefully recieved as usual.

Cheers

Andy