Fitness for service based on minimum thickness

[NovaStark]

Hello all,

I recently took some UT thickness measurements on a water storage tank made up of 1 shell course with a 25 ft diameter (designed to API 650). Based on the measurements, all the lowest readings were obtained in the liquid level region as expected. So when reviewing against the nominal thickness of 6 mm, I get that my lowest thickness is about 3 mm. So over a 10 year service life, that long term corrosion rate is 0.3 mm/yr.

The design calculation via the 1-foot method gives a minimum thickness of around 2 mm. Reading through the code shows that the minimum shell course thickness shall not be less than 4.77 mm for this diameter.

So my question, is a general metal loss FFS assessment required because I am below the 650 required thickness of 4.77 mm or is one not required since I am above the 1-foot method calculated thickness ?

 

[JStephen]

Was the tank lined when built? If so, the corrosion you're seeing may all be in the last year or two, not uniformly distributed over 10 years.

 

[LittleInch]

You really should be using API 653

http://www.api.org/~/media/files/publications/whats new/653_e5 pa.pdf

The FFS spec ASME FFS-1 points you to the 1 foot method, but also part 4 of API 653 which looks to be a much more appropriate code.

3mm thickness looks very thin to me to hold up anything above the liquid level bit I guess if it isn't leaking then you might be Ok

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[NovaStark]

Was the tank lined when built? If so, the corrosion you're seeing may all be in the last year or two, not uniformly distributed over 10 years.

Nothing in the documentation indicates that it was. This is just a receiving tank so I don't expect any odd excursions to cause rapid corrosion.

You really should be using API 653

http://www.api.org/~/media/files/publications/what...

The FFS spec ASME FFS-1 points you to the 1 foot method, but also part 4 of API 653 which looks to be a much more appropriate code.

3mm thickness looks very thin to me to hold up anything above the liquid level bit I guess if it isn't leaking then you might be Ok

My impression after trying to understand 579 metal loss assessment was that it didn't take into account the actual mechanical strength required to hold up the roof and so on. But I'll try using the 653 as a basis. Also there aren't any leaks, so the tank is probably around 7 ft high but the liquid level is about 3 ft where all the low UT readings are. All readings above the liquid level are at or near the nominal thickness.

 

[LittleInch]

API 653 is rather vague when it comes to "other loads" such as your roof load or wind etc and just advises "engineering judgment"(!)

Loss of 50% from the original minimum thickness looks a lot to me like it needs additional support over the corroded area.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[NovaStark]

API 653 is rather vague when it comes to "other loads" such as your roof load or wind etc and just advises "engineering judgment"(!)

After some further reading into 653, it is vague in that sense which is probably why it suggests doing that evaluation along with a storage tank designer. The 579 FFS seems to be doing the same thing as just comparing the averaged remaining thickness against a tmin. Not much to do with the dead and live loads.

Loss of 50% from the original minimum thickness looks a lot to me like it needs additional support over the corroded area.

Are there any special requirements for welding an essential patch around an entire shell course?

 

[LittleInch]

You really need to discuss with a tank vendor as I'm sure this band of corrosion isn't that unusual.

I suspect the issue would be differential tension on the welds as you weld the repair band. Trying to do it with water in won't be easy either.

Can You Drain the tank and then just cut out strips and repair it in sections with some support for the roof?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[NovaStark]

It may be possible during planned downtime.

 

[NovaStark]

By the way LI, I forgot to thank you for assisting me with what seemed to be as something trivial as minimum thicknesses. Currently the only senior people at my company are manager type people, so asking them my question would just yield a response of "you should already know". So thanks again!

 

[LittleInch]

No problem. I just repaired some typos in my post above.

When you repair this tank you might also want to add in a coating layer at the water / air interface to prevent this corrosion occurring again. Or did I mis read this and in fact it is spot corrosion all over in the liquid region?

If you're in that situation it doesn't help when the codes just say use Engineering judgement, but with extra loads on the shell from the roof it's difficult for them to say much more. a competent tank builder / repair company is needed next.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[NovaStark]

The UT measurements were obtained via UT scan which recorded thicknesses at fixed intervals over a specified scan area. The lowest just happens to be about 3 mm which those areas are spread within the scan area.

I probably gave the wrong impression there implying that all the values are 50% wall loss. But there are areas where you get between 4-5 mm and others within the same scanned area where you have regions of 3-4 mm of wall. I should be able to provide a sample of the readings for a second opinion before I go running down a problem that may not exist.

 

[LittleInch]

Ok, I misread your initial post and thought that what you had was a distinct band of corroded material which is common in water tanks which are maintained at a fixed level where the water / air interface creates enhanced corrosion due to highly oxygenated water getting the steel wet and then dry and then wet...

However I now understand that you're talking about much more general clumps of corrosion / metal loss.

Now the calculation in API 653 section 4.3 come into play.

Corrosion though is not an easy beast to tame. Without more definitive results or a series of tests results it's not possible to arrive at a corrosion rate for a particular location. Also at these sorts of thicknesses actual depth can be difficult to be accurate and surface rust can impact on depth readings.

So I think you have a few options
Do nothing and re-test in a years time - Depends on the amount and area of corrosion and your risk profile
Do some patch repairs / painting /internal coating As above, if you're talking about > 15% of the area then you really should go for
Clean and spray / install internal lining over the whole of the tank including the floor

The determination of 2mm probably holds OK as your absolute min thickness, but remember you haven't inspected the floor which can be worse than the walls, especially if there is any level of sediment or "goo" which can harbour interesting bacteria.

Still best to get an input from a tank inspection company / vendor for options and costs, but you seem to have a bit of time to make your decision.

Good luck and let us know what happens.

LI


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[NovaStark]

I uploaded some of the UT data and the B scan view with the thickness scale if you'd like to take a look.

But I opted for a scanned set of readings because as you said, the surface rust would impact on the readings.

We did clean the bottom of one of our bigger tanks as well which uses the same water supply and down there seemed to have a thick layer of slush. But haven't detected any leaks on the bottom plate as yet.

 

[LittleInch]

Wow, that's a lot of data to chunter through and a picture is what you need.

There is a lot of 3.xx readings in there and the picture and data make it look like a lot of long vertical metal loss features.

You've had a good 10 years out of it, probably the only thing you can say with certainty is that you won't get another 10 and if you leave it too long you won't have enough metal to stop it falling down in a big wind

Line it now and it might last another 15.

floor inspection should be undertaken after you clean it - leaking is often difficult to spot on a floor depending on what's under the tank, but is the most common leak point.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[NovaStark]

Wow, that's a lot of data to chunter through and a picture is what you need.

Yeah that's why I also added in the picture from the actual B scan which you can also get from the excel file with conditional formatting.

There is a lot of 3.xx readings in there and the picture and data make it look like a lot of long vertical metal loss features.

Those readings are within the scanned space of the actual equipment used so scans were done at the 4 cardinal points and a scan width of 12 inches with 1 mm spacing per reading. Otherwise I would have tried spot readings but like you said, the surface rust and pitting would affect my thickness readings.

You've had a good 10 years out of it, probably the only thing you can say with certainty is that you won't get another 10 and if you leave it too long you won't have enough metal to stop it falling down in a big wind

Line it now and it might last another 15.

floor inspection should be undertaken after you clean it - leaking is often difficult to spot on a floor depending on what's under the tank, but is the most common leak point.

That's also why I wanted to do an FFS to at least show that while the calculated tmin by the 1-foot method would be quite low, the dead/live loads on the structure may make it not fit for service and require internal inspection. I tried to get my upper people to perform internal inspection on the rest of water tanks (this one is the smallest), but I was mainly met with "well it's just water, so we'll do it another time"

 

[LittleInch]

The fact that it's water is blinding people to the requirements, but all you can probably do it document what you've found to date and if possible get some input from a tank vendor or tank inspection company. Putting off spending money is easy if you're not going to be around after a few years, but when one of the tanks leaks, potentially a big leak and then your production suffers or whatever you use these things for then you can dig out your report that said you recommended some other action.

That's just how it goes sometimes, but your method is still valid and loosing nearly 50% of the metal you start with and using those pictures you have you can also come up with a global metal loss and show it's not just one area, but significant areas of the tank.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[NovaStark]

Thanks for all your help and advice LI. Hopefully my efforts won't be in vain.

 

[NovaStark]

Question about if my recommendation for structural analysis is approved.

Would it better to just ask a company to use something like ametank to design back to code requirements and check or would it be better to ask for FEA analysis and give the dead, live and foundation loads?

I am thinking the FEA is a code type software such as AMETANK might just say that once the thickness is below 5 mm, it fails the analysis.