API 650 - Roof to Shell Junction - 5.10.5.2 4

[Praveence]

​

I have a query regarding API 650 roof to shell junction design.
Tank basic details are, Diameter - 33.5 m, height - 12 m , Self supported - Fire water tank, designed for atmospheric.
Since roof is self-supported and diameter is large, Stiffeners /rafters (external)are provided.

If roof is stiffened by rafters (Externally) do we need to comply to API 650 Cl.5.10.5.2 ?
In my opinion, this requirement is not applicable for stiffened roof but Client is pushing us to comply to this requirement. We do not have any clear reference to defend our claim.

I came across an API discussion on the same topic but not sure how topic was concluded.
Also attached discussion copy which was downloaded from API community, in the discussion same point is raised.
Link

 

[LittleInch]

You should repost this in the storage tank forum under mechanical engineer.

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

 

[JStephen]

I would say the equation in 5.10.5.2 assumes an unstiffened self-supporting roof.
If in your design, you assume that rafters are loaded in compression, then you'd need some way to restrain that compression force at the outside, which would typically be a polygonal or circular tension ring.
If you're using trusses or some other configuration or design assumptions so that there is not a radial component of load at the edge, then you'd go back to the minimum top angle sizes specified elsewhere.

 

[HTURKAK]

-API 650 does not suggest to attach the roof plates of supported cone roofs to the supporting members. As far as I understand, this is also an implicit requirement for the supporting system should be under the roof plates.

- API 650 Cl.5.10.5.2 is for compression ring requirement for the SELF supporting cone roofs. For your design, cone roof with radial rafters still exert radial thrust to compression ring. ( for gravity loading , tension ring ). If you do not want to comply to Cl.5.10.5.2, you may change the concept. Use rafters and girders supported on internal columns, or , use internal radial trusses ( I think better alternative for diameter 33.5 m).

- With your description, the roof supported under radial rafters ,that is, you will need a compression ring at apex of the roof and tension ring at connection to shell.. your plan is not clear for transfer of vertical forces to shell and radial forces ? ..what is your plan for the stability rafters during erection ?

 

[Praveence]

Dear all,
Thank you for your time and response.
1. @LittleInch - I have raised the query in API and mechanical engineer forum for better reach.
2. @JStephen & HTURKAK - This tank is internally coated and as per specification no support is allowed to be welded from inside. Since support is placed from outside so radial thrust will be there.
We are providing a ring at the junction to transfer the vertical load to shell and to weld the roof. My concern is compression area requirement as per API 650 5.10.5.2 which calls for much higher area than what we are proposing. My only concern is do we need to comply to area requirement of Self supported cone roof requirement. If we need to comply then can we take rafter area also in consideration ?

During erection, scaffolding will be used and our contractor has experience of handling this type of tank before.
Vertical load will be transferred through a ring welded to shell as per API 650 Ann.F F-2 - detail d

 

[LittleInch]

praveence,

You should only have one post in whichever forum you see fit.

Ask for the answers to be transferred or cross reference the posts.

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

 

[HTURKAK]

How did you calculate the subject area defined at 5.10.5.2 and the participating area at the roof-to-shell joint determined using Figure F-2 ? .
If the roof type self supporting ( your case , external rafters do not provide exemption ), and if you want to comply with requirements of API 650 , the answer is YES..

The rafters in radial direction which can not be considered for the calculation of tension ring area . The acceptable alternatives are defined at Figure F-2 .

I will suggest the use of internal truss without any internal column. For this alternative, you are free from 5.10.5.2 and will be more stable and economical for this diameter...

 

[Geoff13]

It would make more sense for me to answer under your Storage Tank posting, but I will answer here as there are more replies here.

Your Client is correct, and the 5.10.5.2 area is necessary. In fact it may even be too small.

For a self-supporting cone or dome roof, other than a truss supported roof, the roof will exert vertical and horizontal forces at the roof-to-shell junction. The horizontal forces would be inwards for a tank with sufficient internal pressure to fall with Annex F, but you stated this tank operates at atmospheric pressure, so they are outwards on your tank.

Without the rafters/stiffeners both forces are uniform around the perimeter. This is the scenario that 5.10.5.2 addresses. We can tell this clause is calculating the required area for a tension ring as it uses the allowable tensile strength of the materials. This area resists the outward horizontal component of the load from the roof. The available area follows the rules in Figure F.2.

As you have added external stiffeners to your cone roof the forces will be concentrated, instead of uniform. Thus the Code formula may be inadequate. In addition to the tension calculated by the Code formula there will be bending between the individual load points. You would need to do some calculations to show if these are small or large, and whether additional bending strength is required at your shell-to-roof junction.

You can verify that the Code requires the 5.10.5.2 area for your roof by reading the Note at the beginning of clause 5.10.5 which removes the 5.10.5.1 minimum thickness requirement when stiffeners are added, but does not change the 5.20.5.2 minimum area requirement.

You must also consider if you need to stiffener your tank shell where the concentrated vertical loads will now be imposed by your roof stiffeners.

 

[Praveence]

@Geoff13 : Thank you for detailed explanation. It helped a lot to understand the issue.
I have another query, since load is transferred through rafters, do we need to provide required area as per 5.10.5.2 around entire circumference. Instead calculate (FEA) required area and provide additional stiffening area where rafters are provided.

5.10.5.2 equation assumes load is transferred through plate. In our scenario load is transferred through Rafters and hence compression area calculation may not be correct solution.

Also, we cannot propose internal truss as, client specification clearly states internal supports cannot be provided for internally coated tanks.

 

[Praveence]

Also, I will delete the topic in other threads as good/sufficient amount of explanation is provided here.

 

[HTURKAK]

Mr Geoff13 explanation is excellent to understand the case and a pink star to him for this nice explanation...

-Regarding your first question, I understand that the roof plates will be welded to the lowerflange of the radial rafters. ( is this true ?)..You may perform FEA assuming the roof plates shell elements and the radial rafters and tension rings beam elements. Perform calculation and see the difference with API 5.10.5.2 . The diameter 33.5 m will require primary rafters ( extending to full radius l=16.m), secondary rafters extending from perimeter having l= 12. m extending to radius r=4.75, and tertiary rafters l=8 m from perimeter to radius r=8.75 m.. you will need two polygonal girders supported by primary rafters.The secondary and tertiary rafters will be supported on girders. With this set up, you will see the alternative with external radial rafters is not economical..

-The definition compression area should not mislead.. the compression ring defined with 5.10.5.2 equation, works for compression for internal vacuum pressure and external wind vacuum. The same compression area still necessary for gravity loading and internal positive pressure as tension ring..

-I understand that the client is looking for flat surface interior roof ..Still you may propose internal truss with HSS elements without any internal column.

 

[Geoff13]

The 5.10.5.2 area is the minimum required around the complete circumference.

Let's go back to basics and make a simple free body sketch of the tension ring split across the tank diameter.

Based on the plate span limits of 5.10.4.4 I'll guess you have 50 rafters, and thus there will be 25 radial rafter loads in the free body. I will leave the exercise to you to sum the X component of each load, and then half the result to arrive at the tension in the roof-to-shell junction. However if you get anything different from the 5.10.5.2 result you need to look at your math again.

The tension force in the ring is not the local effect of a single rafter load, but is a global effect of many loads. Thus it must be continuous and uniform even though you are using rafters.

You can do FEA, but it is overkill when a simple free body will answer your question.

I understood that you cannot use interior support like a truss, but wanted to note my comments would not be applicable to that roof support system for anyone reading this thread in the future.

 

[Praveence]

@HTURKAK and @Geoff13 thank you very much for your time. It was very helpful.

 

[HTURKAK]

Dear praveence (Petroleum)(OP),

The tanks having fixed cone roof with external supporting system with a diameter 33 m ( in this case radial rafters ) is not a common case. Not only the final system , but also the assembly of the roof is a challenging job.

If you decide upon final solution with this set up, please keep us involved with your final solution and provide some info, pictures etc. to help, educate young fellows.

Regards..

 

[Praveence]

Dear HTURKAK,
I will definitely update you here. Similar diameter tanks with external rafters are already in working condition at our client site.

 

[Geoff13]

I tried to stay on topic with my answer, but like HTURKAK, I have been wondering about the overall design of the roof. Externally framed at 33mØ would be an interesting challenge (and those are the best kind of projects to work on). If you are willing/able to follow up with pics or sketches that would be appreciated.

Good luck.