Probable Structure of an Existing Water Tank Roof 1

[msquared48]

I have some construction drawings of an existing steel water tank - 60'-8" in diameter - with a domed roof with a 56' radius - apparently all the information available from the local water district. Other than the 8mm (.299") plate for the roof structure, there appears to be no stiffeners used in the roof structure, other than an 1.5" welded lap splice of the triangular roof segments.

From anyone's professional experience with water tanks, is this normal construction for the roof of a water tank with no apparent central column, and using three-dimensional plate theory (snap-through analysis) with a tension tie at the edge?

The intent of the client is to mount eight cell antennas in two sectors on a low space frame on the top, and I am nervous about that - probably no more than 500# total weight. Snow load alone would be about 72K @ 25 psf. I could invoke the no more than 1% live load increase rule of the industry and let it go, but would be happier with more structure here to begin with.

I think that I am going to push the water district for more original structural design information here at the least...

Mike McCann
MMC Engineering

 

[JStephen]

That sounds like fairly common tank construction- roof may be 5/16" rather than 8 mm in the US. Some of the older tanks had structure that was mainly for erection with the roof being stable without consideration of the structure.

Roof design is generally per API-650. In simplest form, roof thickness is crown radius in feet/200, so up to a 62.5' crown radius, you can use 5/16". The center part of the roof may actually be a cone shape. It can be adjusted for loads, etc.

The roof may be an "umbrella" or dome. Appearance is similar, but a true dome has plates that are actually dished to a 3-dimensional shape.

Approach for local loads would be to use the local loading information available for pressure vessels- refer to the various pressure vessel handbooks or WRC 107 or updated version of the same. In particular, try the line-load analysis given in Bednar's handbook, which gives low allowable loads, but is quick and easy. For 500# load, it might involve a stiffener 12" or 18" long welded at the point in question.

For overall loading, just average the new load in with original snow/live loads and use the API-650 approach.

 

[msquared48]

I saw this post on another site and it worries me concerning the current 60 foot diameter of the tank... Bold highlight added. Is the statement true? Remember, this is a domestic/fire water tank...


Re: Roof Structure for Dome Roof Tanks

09/21/2012 1:36 PM



Yusef is correct here..... a domed tank for this pressure and volume is absurd.

We have a cosmic intersection of the inexperienced buffoons.....Both the OP and the ARAMCO tank specifier seem to not understand the limitations of a domed tank construction as well as its advantages and disadvantages.

First, we should not be confusing an aluminum protective geodisic domed cover with a domed roof..... two different things. That seems to be handled above

Secondly, API-650/620 tanks with self supporting roofs (both cone and domed)can only be constructed to a diameter of 10 - 11 meters.....far less than the monster we are talking about here.

Thirdly, there comes an economic point in plant construction where "you simply do not put everything in one container"..... its too expensive ! If the liquid contained here truly has a vapor pressure approaching 14 psig, then several horizontal bullet tanks (or vertical towers with dished heads) should be considered.

Witness the common "farm" of propane bullets, the design pressure is usually 250 psig.... why are many of them used "instead of one big vessel" ?..... because the single large diameter vessel would be too difficult to fabricate, install etc AND TOO EXPENSIVE !

We have a similar design situation here....



Mike McCann
MMC Engineering

 

[msquared48]

Further question:

Considering the statement above, if it is true, could the interior of the water tank be internally pressurized above atmospheric to support the at leas a portion of the live load of the roof structure. I assume the dead load would be self-supporting as the center access hatch removal would relieve any internal supporting pressure. Obviously, this would also impart more external load to the tank sidewalls.

Realistically, this does not seem logical and also seems very costly over time, particularly if the volume of water in the tank is drawn down any significant amount.

Currently I am getting no help from the water district or contractor. I am looking for the designer.

Mike McCann
MMC Engineering

 

[JedClampett]

I can't say what is normal for tank suppliers. But every water tank I've seen (admittedly limited to ground set large diameter) has a flat roof supported by purlins and columns. I've seen 160 ft. diameter prestressed concrete tanks with concrete dome roofs 3 1/2 inches thick. I was sure they were incorrectly designed until I saw the calculations. Needless to say, they were right and I was wrong.
The fact that I've seen so few steel dome roofs (and they've never been suggested for the ones I've designed) tell me they are at a price disadvantage. Maybe forming them or welding them into a shape like that is pretty expensive compared to a flat roof, because it's not just the plate cost. No one every wants to spend one penny more on a water tank than they have to.
I'd talk to a large tank supplier's engineer and pick his/her brain.

 

[SAIL3]

It is difficult to get a good handle on snap-thru loads....as JS mentioned it is probably an umbrella type dome....500# is not a significant load...like 3 or 4 adults standing on it......after I recovered from my "chicken" attack, I would try and address the snap-thru problem as that is where my concern is coming from...if feasible, I would locate and distribute these antennas as close to the edge/perimiter as possible and use some local stiffeners to spread the load out.....these type of projects are never worth it...the risk/reward is too high, especially, if the original design looks like it is marginal...if the OP decides to continue with it, at least, a thorough inspection should be conducted beforehand...

 

[msquared48]

I am going to order a copy of API-650 this afternoon, but in looking on-line, it is hard to quickly isolate any specifics to either confirm or deny the emboldended statement four responses above above.

Can anyone PLEASE verify this as quickly possible as this would constraint be a problem with public safety and structural liability implications here?

I love this s&*((!

Mike McCann
MMC Engineering

 

[Ron]

Mike...if it is a water tank, it was probably designed to AWWA standards, not API.

Bethlehem steel had a steel tank design manual in the early 70's that gave a lot of good info on such designs. I think I still have my copy of it. Will look and copy for you if I can find it.(it was a marketing freebee, so copyright not an issue)

 

[msquared48]

Aaaaahhhhh.. Sooooooooooo...

So maybe the tank roof is OK then. I'll look forward to your post.

Thanks Ron

Mike McCann
MMC Engineering

 

[JStephen]

I've designed and detailed umbrella roofs up to 60-62' diameter, not sure of the maximum.

For a water tank, the main motivation to using an umbrella is lower maintenance due to getting rid of the structure. The umbrella does cost more. It is preferable to build them on the ground and set them in one piece, and as the roof becomes larger and heavier, that becomes more of an issue as well.

The 10-11m limitation above is nonsense. There are practical limits based on cost, so it is uncommon to see one more than 60' or so diameter, but they can be built. Perhaps the person that made the statement was assuming 3/16" plate or something, who knows. But per API-650, thickness t is R/200, R is limited to 80% to 120% of diameter, and t is limited to 1/2". So a 1/2" umbrella can go up to 100' crown radius, and that can be used on a 125' diameter tank. That can be tweaked a bit based on live load used. Larger domes can be built, but they will be stiffened domes.

AWWA D100 references API-650 for umbrella roof design. There is about 1 page in API-650 on umbrella/dome and self-supporting roofs, there's not a lot of detail. They give the equation for thickness, but don't indicate the basis for it. You can derive something similar to the API-650 formula by taking the roof as a free-body, finding the uniform compressive membrane stress at the edge, and equating that to the AWWA D100 formulas for allowable compressive stress in a curved plate/shell, assuming "small" t/R ratio.

Overall snap-through buckling is never considered, as there are limitations in the standard about how flat the roof can be. The design basis seems to be general compressive buckling rather than snap-through buckling.

API-650 also includes requirements for top angle area based on use as a tension ring, but that requirement almost never governs top angle design.

Water tanks are not pressurized to support the roof, and municipal water tanks are not pressurized, period.

I would not anticipate any problem supporting the load indicated, the only issue is showing that by the calculations, and once again, the pressure vessel handbooks are probably the quickest and easiest way to establish that.

 

[SAIL3]

The Op should rely on JS's experience more than on the following concerns that I have....
1. On a pl structure this large I would expect the "as-built" dimensions of the cone roof to vary from
design quite abit which would add another unknown to the pl theory involved.
2. The biggest problem I would have is with unbalanced snow loading on the inverted cone and honestly
would not know how to analyze it on a strucure like this that depends on pl theory alone. If it
had radial stiffeners and a compression ring at the apex , then, I might have a shot at it. Combine
this with the not-well-defined snap-thru condition and I would be really uncomfortable giving this
my blessing.
3. It is not the OP's extra loads thatis the overriding issue here but the existing design and
unbalanced snow loading for which I am not aware of any existing theoretical solution.
4. Just using head/shell theory may work but it is based on a perfect head and uniform loading and is
not what one is dealing with here.
5. Then, again, there are allot of structures out there that are performing adequately but do not
have a clean engineering theory to support their design.
6. Bottom line is I would walk on this one or if, for some reason, that was not possible, I would
attach and extend up a HSS member from the side of the tank and attach the antenna to that. This
way I have not touched the roof and it's current loading.
7. IMO due to climate change the severity and frequency of weather events are increasng and the most
failures recently are due to snow loads on structures that have performed adequately in the past
for 30 or 40 yrs. I am beginning to factor this possibility into my engineering decisions/judgements
and this case would certainly be one of those.

 

[JStephen]

The current API-650 includes provisions for unbalanced snow loading, but that's a fairly new addition to the standard, and a great many tanks are unaffected by the change.

The API-650 self-supporting umbrella and self-supporting cone roof requirements have been in the standard for at least 40 years.

 

[fegenbush]

SAIL3,

1. On a pl structure this large I would expect the "as-built" dimensions of the cone roof to vary from
design quite abit which would add another unknown to the pl theory involved.

The current design process in API 650 is based upon the more stringent analysis of an umbrella roof (plates rolled in one direction only). Additionally, the committee is currently discussing the maximum width of the roof leaves to limit the deviation from the ideal shape.

2. The biggest problem I would have is with unbalanced snow loading on the inverted cone and honestly
would not know how to analyze it on a strucure like this that depends on pl theory alone. If it
had radial stiffeners and a compression ring at the apex , then, I might have a shot at it. Combine
this with the not-well-defined snap-thru condition and I would be really uncomfortable giving this
my blessing.

API 650 5.10.6.1 requires that the design take into account the unbalanced snow load on the roof. I will have to look into older editions to determine if this changed and when. Mike does not specify when this tank was erected.

3. It is not the OP's extra loads thatis the overriding issue here but the existing design and
unbalanced snow loading for which I am not aware of any existing theoretical solution.

See previous comment.

4. Just using head/shell theory may work but it is based on a perfect head and uniform loading and is
not what one is dealing with here.

No comment required.

5. Then, again, there are allot of structures out there that are performing adequately but do not
have a clean engineering theory to support their design.

No comment required.

6. Bottom line is I would walk on this one or if, for some reason, that was not possible, I would
attach and extend up a HSS member from the side of the tank and attach the antenna to that. This
way I have not touched the roof and it's current loading.

You could also determine the allowable point loading based on the ideal shape and limit your load to some percentage of that load to compensate for the less-than-ideal shape.

7. IMO due to climate change the severity and frequency of weather events are increasng and the most
failures recently are due to snow loads on structures that have performed adequately in the past
for 30 or 40 yrs. I am beginning to factor this possibility into my engineering decisions/judgements
and this case would certainly be one of those.

I believe the interjection of climate change into this discussion was unnecessary. The snow loads prescribed in both of the standards mentioned above are based on ASCE 7 maps which have a long history of being balanced between conservatism and economy of structure. If you are bidding a tank, you are welcome to increase the snow load arbitrarily as you see fit. In the meantime, I am going to build tanks to established standards and customer requirements.