ASME RTP-1 - Minimum Thickness of Torispherical Heads 1

[AljosaM]

Hi All,

I'm trying to use the ASME RTP-1 code to calculate the minimum allowable thickness of a torispherical head on an FRP tank. The equation I'm using is 3A-230, it's the first case:

t = (M*P*Rc)/(2*(Su/F))

I'm now unsure of the pressure value, P. The formula doesn't really define it in this paragraph, and I've narrowed it down to two options:
[ul]
[li]Maximum internal pressure (ie. hydrostatic head when tank is full, in my case)[/li]
[li]Maximum external pressure (weight of the nozzles and a 250 lb live load)[/li]
[/ul]

The hydrostatic head (1st) case is obviously the higher value, but intuitively I don't trust it. I see this pressure affecting the hoop stress of the cylinder a lot more than the tank head. The pressure of the liquid would push against the walls of the tank, and expand the vessel in all directions.

However, I believe the stress due to that process would also translate to the tank head.

Can anyone give a little bit more insight on this matter?

Thanks and regards.

Aljosa Mitrovic
Mechanical Engineer

 

[JStephen]

In the pressure vessel world, the thickness would be calculated with a similar formula, assuming interior pressure. (For hydrostatic loading only in a steel top head, that would give a negligible thickness, and actual thickness would be based on code minimums and material availability.)
Vacuum loading would be treated separately with different formulas.
Concentrated loads would be treated separately if needed.

 

[AljosaM]

Thanks for the reply, Stephen.

ASME RTP-1 is specifically written for FRP vessels with pressures up to 15 psi. Meaning that it's not as stringent as the other codes for pressure vessels.

However, my question is more related to the theoretical applications of these given values. If you have any more insight regarding this please let me know.

Thanks.

Aljosa Mitrovic
Mechanical Engineer

 

[EdClymer]

AljosaM

Is this a tank (ventilated to atmosphere); or a vessel (not ventilated and holding internal pressure of liquid or gas (gas over liquid..?) Vertical or horizontal?

I will assume it is vertical.

P is pressure on the head (which can be internal +ve or external -ve). This may have to be derived from external loads; wind, snow, imposed and access for person(s) etc.... and would take into consideration the internal pressure from gas or liquid in contact with inside of head (if any).

It is possible to have two answers of thickness; one for internal (or not) pressure, and one for external pressure.

3A-230 calculates minimum thickness of head - but I would expect code minimum thickness to apply if that value is greater...(all of the GRP (FRP) codes have a minimum thicknesses).

In the case of vertical cylindrical tank or vessel; the static head pressure has no loading or effect to the shell in axial (length/height) direction. All the loading from this static head is circumferential (aka hoop). In a ventilated tank, circumferential/hoop stress is always much greater and is the decisive loading.

If you really want a struggle; try the new(ish) (5th attempt) EU code EN13121-3.

Trust this is of help.

Regards,

Ed










Ed Clymer
Resinfab & Associates
England

 

[AljosaM]

Ed,

That's exactly what I was asking; I think we're really close to a solution.

To give more insight, I am dealing with a vertical tank which is open to the atmosphere. The external forces can be considered negligible (for now), and I plan on using a separate equation to derive the loading due to earthquake/wind effects.

I'm interested in what you said about the effects of the static head when calculating the thickness of the of the top shell of the tank. See the attached sketch I prepared:

As you know, when you fill your tank with liquid you will have some forces acting on the walls of the vessel which will translate to hoop stress. If we assume a certain degree of elasticity in the tank walls, these forces will also translate to the junction between the torispherical head and the cylindrical wall.

My question is, what do we assume the pressure (or force) acting upon this junction? It seems a little excessive to use the entire hydrostatic head, but maybe it makes sense.

Hopefully the sketch helps. Let me know what you think, and thanks again for your reply - it's a big help.

Regards.

Aljosa Mitrovic
Mechanical Engineer

 

[EdClymer]

Aljosa

I have looked at sketch; worth many words.

Item is a tank (because it is ventilated) and pressure greater than atmospheric can’t exist above the liquid. I also note there is no overflow branch, so, in theory the head can fill with fluid until it overflows. The question you ask is; what force can be assumed at the head to shell joint…best not to assume - but calculate.

Loads on the head - externally; will be self/gravity weight; snow and wind; superimposed load (men and toolboxes) and any other feasible loads. These are converted to effective pressure (which are likely to act together I will leave to another day).

What happens if someone, due to odour from tank (when it exhales) decides to connect vent branch to ducting and a fan giving -ve pressure - good for replacement heads. Such heads are vulnerable to external pressure.

If the head fills with liquid (or you decide it could), that will give a (small) static head at the head/shell junction. That could be used to calculate thickness - but minimum code thickness will over-ride. From minimum code thickness (or calculated thickness) you can work backwards via laminate properties at that thickness - then what pressure at shell diameter is required to give that laminate loading. You will now have a pressure.

The head/shell junction will require a thicker section, treat as a knuckle and factor loading x 2 - 2/3rds outside, 1/3rd inside over length as calculated.

(Now run out of space...will send part 2 later)

Trust this helps.

Regards




Ed Clymer
Resinfab & Associates
England

 

[EdClymer]

Aljosa

(Part 2)

If head pressure is really zero, then no thickness, no knuckle, and nothing to stop head falling off.

I forgot seismic effect, where liquid accelerates upwards and then collides with inside of head..so better take that into account and correctly design joint head to shell. In this case the force on the head will have to be converted to a pressure.

As to the question of liquid in shell. When emptying or filling, the liquid moves independently of shell. If shell diameter changes then height of fluid changes. The static head pressure has no loading or effect on the shell in axial (length/height) direction. All the loading from static head is circumferential (aka hoop). In a ventilated tank, circumferential/hoop stress is always much greater and is the decisive loading. But, laminate which may satisfy circumferential loading may not satisfy axial loading. Take worst case and use that. I think that your code does a combination check on circumferential stress when added to axial stress etc has to be <1.

Again, I trust this helps; if you have any more questions I will try to answer them.

Regards

Ed






Ed Clymer
Resinfab & Associates
England