Tank shell seam weld corrosion - what to do!

[robsalv]

Not entirely sure this is the right forum, so happy to take suggestions on that.

We're a downstream petrochemical plant in Australia, in a self regulated environment. We have an open topped, API12C, million gallon CS water tank as the main water supply to our firewater and plant "process water" pumps. Due to shell corrosion at the normal liquid level, the tank had to come out of service for repairs - its first time in 46yrs. This took quite some doing!!

Inspections to date had been via external UT surveys, which did not include angle scanning of the weld seams... in hindsight, that was quite some oversight!

One of the surprise findings was the extent of preferential corrosion of the shell weld seams - all of which have corroded to some degree - up to 30% strake thickness in some areas. Interestingly, the floor to shell weld is still pristine and most shell plate has general light pitting only.

Anyway, according to API653, weld seams need to comply with the asbuilt standard, which today is API650. API650 says that welds need to be full pen and full thickness... which is not a very fitness for service type of answer IMHO.

My question is this - must I repair near on 2kms of weld seams?? [$budget estimates $200k+]

Or...

Since many weld seams areas are still thicker than the locally thin minimum thickness calculation of API653 clause 4.3.3.1., can I pursue a "clean up, blend and stripe coat" technique to eliminate future corrosion of these particular weld seam areas, thereby reducing the scope of weld repair?? Am I within the spirit of the codes doing this??

Appreciate your thoughts.

Thanks.

Rob

 

[metengr]

Look at API 579, as well. You can certainly perform a risk assessment and decide on a course of action.

 

[robsalv]

Thanks Metengr... thing is, how does one apply API 579 in this case. The situation is not a local crack like flaw or bit of weld LOF, nor a discrete gouge, nor an LTA...

Surely there's a "residual strength" remaining in the welds, but I'm at a loss to see how to utilise the API579 tool.

Appendix A, clause A.6 API650 Storage tanks says:

The equations to evaluate the minimum thickness and maximum fillheight of an atmospheric storage tank are covered in section 2 of API653.

The reference to section 2 MUST be a typo, since those equations are in section 4... (interestingly, this is a consistent typo, as section 2 is referred to in any tank related example calculations throughout API579)

I'm not an expert in API579, but it seems to me that the general metal loss and locally thin area sections, 4 and 5, are tools for analysing losses away from major structural discontinuities, i.e., pressure containing welds.

Any further thoughts??

 

[metengr]

robsalv;

One of the surprise findings was the extent of preferential corrosion of the shell weld seams - all of which have corroded to some degree - up to 30% strake thickness in some areas.

I would blend grind (clean-up) the corroded area of the shell course welds to simply reduce the effect of a crevice which would also facilitate NDT, and then treat it as an LTA. I doubt you would have to do anything beyond running some LTA calcs to assure yourself of a sanity check.

 

[JCAT96]

If the welds are still above the minimum required plate thickness then I would just clean them up and blend them in. If however they are below the minimum thickness I suggest you will need to build them up again. The vertical welds are the most important in this regard. The Horizontal welds can be treated as a LTA if the thinning is over relativly short distances.
You will need to look at how you monitor the tank in future with regards shell thickness at the welds!

 

[robsalv]

Thanks for the replies folks.

On the topic of LTA's, we're talking weld seams, so any blend grinding would end up with grooves of about 0.5" - 0.75" in width. The depth would be up to the minimum required thickness per API653 4.3.3.1b) with the radius in some locations quite small (doube V, square butt type weld) since the welds would need to be chased/blended to remove most of the honeycombing. That leaves me needing to assess using API579 section 5, groove assessment. My stumbling block with this becomes the length of the groove.

A groove might be ok in isolation for a short length, but API579 isn't set up to treat extended grooves, and the issue is that the entire tank's shell welding is impacted.

Short of an FEA solution, I may need to go with the reweld option and that's big unplanned biccies...

If welding is the pathway I finally recommend to management and wear the heat, then I'm thinking of an arc gouge and flux core type welding spec, basically to speed up the process and reduce the cost.


For your edification, some examples of the corroded weld seams...

Some pics:

 

[metengr]

robsalv;
Despite the rather unattractive locally corroded surfaces, since this is an atmosphere water storage tank, you might be surprised as to how much design margin there is with the remaining wall thickness. In lieu of additional welding, you might be able to get buy with blend grinding the corroded locations and coating the interior surfaces to preserve the remaining service life of the tank.

 

[StoneCold]

After 45 years of service you really might be interested in a replacement tank. Even if you dump $200k into it you still have an old tank that is probably only a few years from leaking somewhere else. Maybe you could put a bladder inside of the existing tank and forget about it?

Just a thought.

Regards
StoneCold

 

[robsalv]

A bladder has been considered, as has a new tank (which installed is upwards of $4M Australian.)... but it's clear we have some structural issues that need dealing with anyway, so the tank is still slated for bringing up to service fitness.


Close investigation of the lower strake welds show greater loss than first thought. API579 groove based LTA calcs say that the bottom strake welds need repair... the welds on the higher strakes though look to be blendable and coatable... I knew there was middle road.


Regarding the "if in doubt chuck it out" philosophy, if the tank's fitness for service can be maintained economically, why would I recommend scrapping?? I'm confident my inspectors would be able to locate the areas that are "a few years from leaking" to allow them to be brought up to service fitness.

Cheers

Rob

 

[MJCronin]

rob....

Aside from love, nothing lasts forever.......you got 45 years service out of the tank,........ many bridges, buildings, roads and people don't last that long.

I would replace the tank and incorporate some of the new High-tech coatings systems into the tank design..... If this tanks supplies process water, why was an "open-top" design used ?....It seems to me that this would only serve to promote corrosion...

Please complete the thread and let us know your final decision..

regards

MJC

 

[robsalv]

...well, the estimate for remediation repairs keeps going up and up... so I'm putting the options in a report to management and we'll see where they want to go...

Remediation to buy 10yr life is in the order of 20% of total capital replacement cost.

Keep you posted.

If there are any Aussies on the thread, there's a better than even chance this tank might make it on an experience paper for the 2007 Oil and gas symposium.

Cheers

Rob