Vacuum

[parrot77]

I am currently competing a report on a roofed thin walled steel tank, whose walls have structurally failed. The designers calculations only designed for the load condition whereby the tank was full, ie hydrostatic presssure. The tank was being emptied for servicing however, through a scour valve at the bottom. Negative pressures or a vacuum built up over the head of liquid as the liquid drained away. As the level of liquid in the tank reduced the negatve pressure caused compression buckling in the tank walls. Can anyone indicate to me how i would calculate the vacuum presssure in the sealed tank? How would this vacuum pressure manifest is as loading on the wall over the liquid?

 

[JedClampett]

If the tank pulled a total vacuum, the pressure perpendicular to all external surfaces is atmospheric pressure.

 

[JedClampett]

I looked it up. It's about 15 psi.

http://en.wikipedia.org/wiki/Atmospheric_pressure

 

[PatBethea]

In the past on this forum, I've seen posters (myself included) directed to "hire a structural engineer". Let me return the favor - Hire a mechanical engineer. At the very least, you might want to repost on the Boiler & Pressure Vessel Forum.

Jed is correct - at SEA LEVEL, the external presure acting on the tank is approximately 14.7 psi. If you're in Denver, you might want to adjust that figure a bit.

How certain are you that the collapse was a result of vacuum? Are you sure the tank wasn't being vented during the draining operation? Depending on the tank's size, age, materials of construction, and/or service, it could have collapsed upon itself once the liquid was removed, i.e. it couldn't support its own weight.

 

[DonPhillips]

Keep in mind the pressure inside the tank probably never reaches 0 so the walls probably never sees the full atmospheric. Perhaps you want to calculate what the failure pressure would be and if it much less than 14.7 psi, then you at least confirmed the tank was not designed to be under a vacuum.


Don Phillips

http://worthingtonengineering.com

 

[PatBethea]

Don, don't get too far ahead of yourself. We've been told it is a "thin walled steel tank" and that it contained a "liquid". There are a lot of unknowns - size, age, liquid, operating parameters. It could be there's been significant corrosion and as a result, the tank collapsed under its own weight once the "liquid" was removed. So, unless the OP has some measurements on the actual wall thickness, a calculation as you described wouldn't tell you much.

From experience, it is doubtful that vacuum was a design consideration - only if the vessel is to operate under a vacuum is it considered. I still find it hard to believe that the tank wasn't vented at drawdown.

 

[dig1]

There are a variety of factors which determine the amount of external pressure which can develop in a tank/vessel during draining. PatBethea is correct about hiring a mechanical engineer or at least get someone involved in tank and vessel design this is not a simple problem.

Someplace I have, or at least used to have, a spreadsheet where we looked at this calculation. IF I can find it I'll let you know. Some of the issues are how fast the tank was drained and the size of any vents. Also, was the tank empty (or nearly so) when it collapsed?

Thin walled tanks generally can withstand very little external pressure (negative internal gauge pressure) and therefore almost all atmospheric tanks have either a vent, overflow (which can act as a vent unless it is piped to the bottom of the u-drain which the tank is being emptied into and then it cannot draw air back into the tank and cause a collapse (seen that done before)), or other means to allow air into the tank. Due to the viscoity differences between air and water, the air opening does not have to be extremely large compared to the drain.

Calculating the failure pressure is going to be a relatively inexact exercise. The pressure acting on the roof places the side walls in axial compression plus the external hoop pressure. Then considering the inherent out-of-roundness and other geometric imprefections plus any shell loss due to corrosion and the problem becomes more difficult. Failure analysis have to be looked at differently than design problems because you need to strip away the safety factors and understand how the structure behaved during the failure.

Patrick

 

[parrot77]

The tank is 12m high and contains wastewater so not sure that it qualifies as a boiler!!! Yes, I had thought that for equilibrium the wall would have to resist atmospheric pressure if it was not vented. However, the loads involved seemed so high seemed that I was unsure. The tank was relatively new and in good condition so I do not believe it collapsed under its own weight. I will look more closely at the arrangements made to vent the pressure.
Thanks for all your replies guys!!!

 

[rmw]

Parrot, our (former) tank isn't a boiler it is a (poorly designed IMHO) pressure vessel that has failed most probably (based on no more info than you have given) because it was never designed to go under vacuum, much less full vacuum.

The Boiler and Pressure Vessel forum is more about pressure vessels and boilers as pressure vessels than boilers as operating steam generators, although the topics are well mixed among the threads.

The recommendation to post there was a good one because the folks that lurk there are all about pressure vessels and probably probably more so than they are in structures. If you read through the treads there you will often see the term "FV" mentioned there which is a common tank and pressure vessel term relating to the vacuum design rating of a tank or PV.

It would appear from what is given that your tank was never designed for full vacuum or even much vacuum and somehow it got there and as such was an accident waiting for a place to happen. If you have the original design specs, that is the place to start looking.

rmw

 

[parrot77]

Grand,
Thanks rmw, I'll do that. Ye I dont recognise some of the abbreviations/ units because I am practicing in Ireland. So I'm more use to kN/m sq and kpa!! TANKS again (excuse the pun!!)

 

[40818]

Just a thought, if you say you suspect the failure was because of a pressure differential of 14.7psi, then you would expect the tank to fail (starting from the bottom of the tank) also under the internal pressure due to the waste water if the pressure at the bottom of the tank were greater than 29.4psi
You say that the tank is 12 meters high, your pressure at the bottom of the tank (P_total = P_atmos + P_fluid) will probably exceed 29.4psi

 

[hokie66]

404818,

The pressure differential due to 12 metres of water is about 17 psi, but that puts the walls of the tank into tension. A vacuum situation, if that is what it was, would put the walls into compression, a much worse condition.

 

[rmw]

Parrot77,

Sorry,

IMHO = In my humble opinion
FV = Full Vacuum
PV = Pressure Vessel.

I loved your pun. You are welcome.

rmw

 

[40818]

hokie66.

If the tank walls are sufficient to accomodate bending, then one side is tension and the other is compression, regardless of whether the tank is being expanded or contracted. If however, the tank walls are very thin compared to its dimensions, then it would form into a membrane tension effect, even if the air inside is being evacuated (though stiffeners will help)
If your right with the pdiff of 17psi, that value is still above the 14.7psi from atmosphere which is being touted as the vacuum pdiff.

 

[DarthSoilsGuy]

what purpose does the tank serve? is it for temporary storage and is expected to fluctuate?

any chance, it had it been recently inspected? i'm not sure if EPA SPCC legislation applies to WWTPs, but i have inspected community WWTP's ASTs with the same methods. Is the tank insulated (limits effectiveness of inspections)?

any chance that there may have been concentrated grit scour and the associated accellerated corrosion from the wastewater attacking an interior portion of the sidewall.

i've performed steel tank inspections with ultrasonic testing and have found this to be common with wastewater tanks.

too late to take a good look at the roof at this point i suppose. but, i would be looking to see if the roof has changed since design, with more isolated (& eccentric)loads, or if the roof system is a built-up type where you could have unknown load distributions.

another common item i've found working with steel tanks for wastewater, the area right above the water is usually the worse (unless there is a grit scour attack).

you should also check to see if any modifications/retrofits have been done for service reasons, if you haven't already.

 

[parrot77]

40818,
The tank walls are certainly not thick enough to accomadate compression / tension zones assocaited with bending action. Instead when the tank is full the walls will develope hoop tension stresses to contain the liquid. As hokie66 said the compression resistance of the thin walled section will be very small and the failure was all the way round the tank unlike any concentrated corrosion as suggested. I am definitely leaning towards the theory that venting was inadequate. I will invesigate further and let you guys know what I dig up!!!

 

[hokie66]

40818,

What I assumed, but which was not explicit, was that the tank was round, taking all the load in hoop tension. Don't you guys design airplanes that way?

 

[40818]

I was thinking the tank was a square/rectangle.....
And yes, hoop tension is encountered.

 

[rb1957]

it was a dumb a$$ design not to include a vent.

actually, without a vent wouldn't the original contains of the tank (ie the air inside in) be compressed when it fills, and so expand again as emptied ?

unless of course it was filled upside down, so the air inside could escape ... (not!)

seriously, tho', how is the tank (sorry, ex-tank) plumbed ? is there a delivery inlet at the top (which would have allowed the original air to escape) and which would have been closed to empty the tank ?

 

[JLNJ]

We recently had a failure of a stainless steel non-PV mix tank which crumpled like a stepped-on soda can when it was emptied without the vent opened up. It was ugly.