tank roof uplift

[sme75]

Is it typical to dsign for an uplift, due to wind, on the roof of storage tanks? API 650 and AWWA D-100 do not seem to address the topic, they just give a formula for the lateral load. In fact the anchorage calculations in section 3 of API 650 do not include an uplift due to wind in table 3-21b. If you have used an uplift design force, what have you used to calculate the force? Thanks for any ideas.

 

[JStephen]

It is not typical to design tanks for an uplift for wind. API-650, AWWA D100 and even ASCE-7 all give wind loads based on projected area, so you get a wind overturning moment, but not a vertical wind uplift.

 

[IFRs]

If the fixed roof has sufficient slope, like an aluminum dome roof, the dome roof structure is analyzed for loading including uplift. Stresses from the uplift are contained within the dome roof and not transmitted to the tank other than vertical uplift. The uplift of the dome would tend to offload the tank structure, making it a non-critical load case and is safely ignored for the tank design.

 

[sme75]

IFRs,

In this case the roof is a 15 degree cone with external stiffeners. Do you feel that coefficients for a domed roof would apply in this situation? As JStephen mentioned, API, AWWA, and ASCE all give lateral design forces, but do not address uplift on the roof of tanks. The question originated from a request to check the sliding of the tank under wind loads, including an uplift on the roof. With a large tank, an uplift on the roof will quickly reduce the allowable friction force at the base. This will make uplift on the roof a critical load case if there is not enough tank weight to resist the uplift and lateral wind loads.

 

[UKCats]

If designing a tank to Appendix F (Small Internal Pressures), shouldn't uplift be considered since both the internal pressure, uplift and bending stress are trying to separate the tank from the bottom plate???

Also, I was told that in the older codes didn't differentate between cone and domed roof uplift coefficient. My doom roof on a 1972 existing tank has a uplift coefficient 3x of a cone roof and it fails the 1.5 in w.c. rating of the tank. It passes with a cone roof. Anybody know the history of uplift coefficients on storage tanks.

 

[IFRs]

I'm a bit out of my expertise here, but I find this conversation stimulating and wish to see it continue.

1) Would the external stiffeners disturb the flow of wind over the roof and cause enough turbulence to at least introduce doubt as to the reliability of the pressure coefficients?
2) If your analysis shows that the tank will slide off the foundation, won't anchors take care of it?
3) How do you determine the friction of the bottom on the foundation? Is the tank bottom cone taken into account?
4) If the uplift exceeds the weight of the cone roof and it's structure then the outer edge of the roof and tank may go into compression, leading to a different failure mode.
5) Can the owner put liquid into the tank if a hurricane approaches? Even water? There is usually plenty of warning.
6) What is the diemater and height of the tank?

 

[JStephen]

I'm assuming that the tank isn't pressurized. If it is, you certainly have pressure uplift, but that isn't wind uplift.

Keep things in perspective here. Most tanks are operated somewhere between totally full and totally empty. The sliding-in-the-wind problem assumes that the tank is just totally empty, so it's an uncommon condition to begin with. I can't think of why you'd want to approach this particular design aspect in an extra-conservative manner.

As to the wind on the roof with external stiffeners- using the API/AWWA wind loadings, it would probably make sense to consider the stiffened roof as "flat" area at 30 PSF or so instead of coned/double curved at 15 PSF.

Determining the friction is easy- it's stated in one or both codes (I think it's the tangent of 30 degrees, but would have to check). Testing it is easy- take a steel plate, a chunk of concrete, tilt the plate till the concrete slides and measure the angle- that'll give you the coefficienty of friction.

Unless standards say otherwise, I'd include the floor weight in the resisting weight for sliding- I can't imagine that the tank's going to slide off and leave the floor there.

 

[sme75]

I agree that this is overly conservative and that the tank is not going to go anywhere, however I can't seem to convince the foundation designer/customer that this is the case. They want to see sliding calculations for an empty tank with wind uplift on the roof. The tanks are anchored for overturning with bolt chairs, so if the tank moves the bolts will bend and I believe that that is the real concern that they have. So this leads to my original question of what would you use to calculate the uplift on the roof? I don't think the codes address it and have not seen it calculated in other storage tank texts. I also wonder if the stifferners, which are at most spaced at 6.25 feet, would disrupt the flow over the roof like IFRs said. Using an uplift number they provided all but one tank checks out for sliding, but in addition to this I will also need to re-check the top angle ring for the uplift/compression in the ring. Do you know of a reference that I could use to clear this up for them?

 

[TankDude]

The resulting uplift from wind load is simply the result of the loading from lateral wind pressure striking the projected area of the tank, roof included. You have to find the projected area of the roof slope and multiply it by the applicable shape factor from API 650 and the wind pressure. You then multiply by the distance between the centroid of the roof and the tank bottom to get the contributing overturning moment. Along with the o.t. moment from the shell, the uplift case is listed in API 650 Table 3-21b (4th down). To say there is true uplift of the roof due to wind is inaccurate (provided the roof to shell joint is properly sealed). App. F does not apply for wind load (i.e. external pressure). Bear in mind that if this is a self-supported cone roof you can contribute the entire weight of the roof plate to the overturning resistance (as opposed to a structural roof). The sliding question comes-up all the time and while not proposterous, it is still over conservative to consider this as the only reason to anchor a tank. If the tank is already anchored then that should be more than sufficient to stop the sliding due to wind.

 

[sme75]

Tankdude,

I have already done what you describe and this is not what I am asking. I am asking if there is any requirement to design the roof for a vertical uplift force due to wind. This would act similar to an internal pressure and lift up on the roof. This is typical in building design and the ASCE 7 code addresses it for buildings. I don't think it addresses this for tanks and this is further supported by the lack of coverage of this topic in the API code. It seems that from the responses that are coming, the answer is that this type of uplift would be atypical for a storage tank design.

 

[TankDude]

sme75,
I would have to say that you're right, it's atypical for a tank design. I am not very familiar with building design, but I cannot see how there can be any uplift from wind when there is virtually no projected area on the underside of the roof for the wind to affect. I will look through my ASCE 7 to see if any of the loading cases seem appropriate. Otherwise, I think you have designed the tank correctly.

 

[JStephen]

"I am asking if there is any requirement to design the roof for a vertical uplift force due to wind." The answer is "No."

"The tanks are anchored for overturning with bolt chairs, so if the tank moves the bolts will bend and I believe that that is the real concern that they have." The bolts should have some nominal amount of pretension, and that force can be added to the weight in figuring the friction- that's how the bolts resist sliding. If they don't have any nominal amount of pretension, add it- it shouldn't take much. If you needed to, you could add a welded collar at the base of each bolt chair specifically for shear loading, but this shouldn't be necessary.

You might note how they work this at a shear plane in concrete in ACI-318- they figure the rebar will deflect slightly and then develop its full tensile capacity across the joint. It's obviously not a load situation that you'd want to reverse repeatedly.

 

[UKCats]

Isn't the uplift force due to the aerodynamic shape of the roof. Wind blowing across a foil will cause lift. A dome roof has a higher uplift coefficient than a cone roof. I have a graph that shows the uplift coeff is 3x on a dome roof than on a cone roof.

 

[JoeTank]

When wind blows across a cone or dome roof there is an uplift force. Several wind tunnel studies have shown this behavior. The uplift coefficents are higher at the windward edge of the tank and minimal at the leeward side. The current API-650 does not require that this load be addressed for cone roofs. The unfortunate problem is that when one applies building code rules to tanks strange things happen on paper. Tha last time I saw this done it resulted in a 50 ft diamter water tank having anchorage at about every 5 feet around perimeter. Really looked odd, but owner's consultant was convinced that the tank would uplift and drift away in a 100mph wind. The uplift forces have reared thir ugly heads a few times. Hurricane Andrew ripped the roof off of a few tanks in Florida.

Steve Braune
Tank Industry Consultants

http://www.tankindustry.com

 

[UKCats]

A saw a study from Huricane Andrew that reported a tank's domed roof failed at the top seam because of uplift and actually prevented the shell/bottom seam from failing. The weight of the tank's contents kept the tank from blowing over per se.

When did the wind tunnel study of dome roofs occur. I heard in the mid to late 1990s the increased aerodynamic affects of dome roofs entered into standards.

 

[JStephen]

I would suspect that even the tanks damaged in Andrew were not designed for the wind velocities encountered.

 

[IFRs]

API 650 Addendum 4 dated Dec 2005 and just released for purchase includes uplift due to wind as a required load. If I read it correcly (and I just got my copy today), the uplift is 30psf on horizontal projected areas of conical or doubly curved surfaces. This is for a wind speed of 120mph and is raised or lowered by the ratio of the velocities squared. They allow an alternate calculation by ASCE 7.

 

[sme75]

Thanks for the discussion. I will look into the latest addendum of API 650, I have been waiting for it to come out to at least close the gap between ASCE and it on the seismic loads. It looks like they have also addressed this issue as well

 

[JStephen]

Hey, thanks for mentioning that it is out- been expecting it for the last year, I'll check into it.

Is that uplift required on all tanks? It sounds high.

 

[UKCats]

I believe it also states 18 psf on vertical cylindrical surfaces, corrected for wind (V/120)^2.

It also says 30 psf for horizontal projected areas of conical or doubly curved surfaces.

Not being a structual engineer, does that mean you use 18 psf for overturning moment (vertical projection) and uplift I hope and 30 psf for compression ring (horizontal projection)???