Storage Tank Welding Procedure

[Leinster]

I would like to ask for some advice on welding sequence on API 650 storage tanks which I know is not covered in the standard.

When erecting a storage tank with a crane we have typically fully welded the verticals of one shell course before welding the verticals of the course above and then the horizontals between the 2 courses to allow free movement of the shells due to heat to minimise on residual stresses.

Our site has put in a request to not weld the entire external joints of all of the shells (vertical and horizontal) and then come to the inside and weld all of the shell joints after. The reason being that the tank can be built far quicker this way. I am a little concerned of the residual stress in the shell plate when it comes to welding inside as none of the plates will be free to move.

Some advice would be highly appreciated.

 

[Duwe6]

API-650 tanks are *very* flexible, and 650 is a "Standard", not a "Code". Thus nobody much works to the inordinately strict recommendations, nor do they demand a revision to allow industry standard practices to be used.

IMHO:

1) Completing the outer ring of floor plates , or annular ring, must be done prior to welding the 1st Shell Course "tub ring" to the floor.

2) The 1st course "tub ring" vert's must be fully complete prior to welding the 1st circ weld -- Shell-1 to Shell-2. Nor should any tacks over about 1-inche long be allowed on this circ joint. If big tacks are needed for stability, like erecting in the wind, they must be removed prior to starting the 'production welding' of the S-1 to S-2 root.

3) Assuming a 'normal' thickness of 5/16-inch or less for S-2, one side of the S-2 verts must be welded prior to 'production' welding of the S-1 to S-2 circumferential seam. Reason is that 'Tankies' typically use a large E-6010 electrode to weld shell seams with. With a modicum of skill, the 1st-side [typically the OD] weld will be 1-pass and achieve approximately Full-Penn of that half of the double-welded [OD, then ID] full-joint. At this point, most of the shrinkage, and thus the potential stress, is done.

This last recommendation of mine is in direct contravention to API-650, but is 'usual & customary' for field tank erection in USA. It will give you an excellent tank. It has for me on tanks ranging from 15-ft to 125-ft in diameter, of NORMAL thickness. Meaning the shell is no more than 5/16" thick. For thick -- 1/2-inch, 12mm -- tanks, I would follow the API sequence exactly. My personal opinion is that the writers of API-650 intended the erection welding sequence to be "One size fits All". IMHO, it is excessivly limiting when used on 'regular' tanks.

If this sequence would leave any significant residual stress on the 1st circ sean, it would show itself by this seam exhibiting "positive Banding". The circ seam would be creased in, or peaked out, of a straight & plumb vertical line -- Banding. Since the final welding was performed on the ID, the OD would be bent outwards ['positive' direction].
If you don't have significant Banding, you have very little weld stress. This is proven by the small ammount of weld shrinkage [which causes the Banding].

your mileage may vary

 

[Leinster]

Duwe6...

Many thanks for your informative, very helpful answer.

So essentially if the distortion (deviation) is within the limits set by API 650 then it has to be assumed that the residual stress in the shell plates would be considered acceptable?

 

[Duwe6]

Correct. Thin -- 1/4 to 3/8" -- plates will distort due to weld shrinkage, and thus relieve the bulk of the weld-shrinkage induced stress. "If it looks good, it is good"

Do NOT use this relaxed erection/welding on anything thicker than 3/8". It would probably work up to about 1/2", depending on the size & thickness, and direction of each weld bead deposited. Vertical-up is better than 'downhill' and thick beads are more stress-relieving than small, thin ones. But don't gamble the integrety of your tank on my "probably". Only relax the sequence at 3/8" or thinner. And be prepared to justify your reasoning.

 

[Leinster]

Duwe6...

Many thanks again and one final question.

You have stated that "If it looks good, it is good" is only applicable to thinner plates but not plates above 1/2". Why is that the case? Could residual stresses still be in these plates even though there is no significant distortion?

 

[Duwe6]

On the thicker plates, you can [will] have higher residual stresses because the thick plate will better resist the weld shrinkage forces, leaving residual stress.

There will ALWAYS be residual stress in the welds, any welds. The question is "Is the residual stress low enough to be acceptable?"

Distortion relieves weld stress
PWHT essentially eliminates weld stress
Large thick weld beads 'temper bead technique' will heat, partially remelt, and thus relieve the stress on the bead immediatly beneath them.

So, as you suspect, the residual stress on thick plates may be unacceptably high, *especially* if there is no significant distortion. Distortion relieves stress. Your stress level just needs to be low enough that the product stored will not cause Stress Corrosion Cracking -- SCC.