I'm trying the new TANK "wind profile" function, that defines the wind pressure at different heights.

I am used to use this feature in PvElite too, and pvelite clearly states that the user should input pressure values already factored with the correct shape factor (0.7 or 0.8 or whatever the designer prefers).


This means that pvelite will NOT correct the wind profile and it will use it "as is".

Now, going back to TANK, there is no similar hint in the "wind profile" window.
I assumed that, TANK would require factored pressure too, just like PvElite, but in fact the program is not stating this, and the user manual too..
The point is that I have a spreadsheet which gives me the overturning from a given geometry and a given wind pressure, and if I cross-check with TANK results, the resulting overturning calculated by TANK is way lower (40% - 50%) than the one calculated with the spreadsheet.

I have the doubt that TANK is not requiring a factored wind profile, but is requiring a "pure" wind pressure without any shape coefficient. Then, during analysis, the wind profile will be factored again.
Is this what the software is doing?
Otherwise, I can't find an explanation to understand why Tank is giving me a wind overturning which is lower than the spreadsheet overturning (the obvious explanation would be that the spreadsheet is wrong :P ... I am checking this too, but it has been used for a while now, without major objection by customers and fabricators...)

Hello Fedeghi

I ran a sample tank in the programme, and it appeared the shape factor or any other factors had not been used to modify the user defined wind pressures. The pressure are used un-modified.

This is not an answer to your question-I'll just try to explain the background if API 650 "wind pressure" on shell.
I think it is useful because offer you the possibility to compare with other approach.

The API 650 design pressure is based on ASCE 7 and API 650 details the calculated pressure in 5.9.7.1 Note 2 which gives the value of 31.0 psf.

As explained in footnote (b) of 5.9.7.1, the pressure is assumed to be uniformly applied over a local area of the tank shell (equal to the theoretical buckling mode), and a shape factor is not applied. API 650 simply says "b. The wind pressure is uniform over the theoretical buckling mode of the tank shell, which eliminates the need for a shape factor for the wind loading." based on a distribution of wind pressure consistent with the buckling mode.

ASCE 7 gives a force coefficient (Cf ) of 0.5 for most API tanks- see ASCE 7 Figure 29.5-1 for "moderately smooth surfaces, a height-to-diameter (h/D) ratio of 1 and D*sqrt(qz)> 2.5). Note that Cf is 0.6 for (h/D) ratio of 7, so the coefficient Cf is near 0.5 for any tank.
This coefficient is applied to a design wind pressure of 31 psf, this yields an average pressure on the tank shell of 15.5 psf but API 650 conservatively left the shell pressure at 18 psf which is equivalent with Cf=0.58.

My best regards.

I would add some remarks on the terminology "shape factor".

Clearly, in Fluid Mechanics and in modern codes there is a difference between the "pressure coefficient" and the "force coefficient" (or "drag coefficient").

In a more general case, a pressure coefficient is related to Reynolds number, surface shape and point location on the surface considered. If the subject is the the wind action on tank shell, a pressure coefficient is related to pressure distribution over the cylindrical tank surface and we may note that it slightly depends also on Re number.
Such pressure coefficients have been included in standards- see for example the old BS 6399 Part 2 where Table 7 details "external pressure coefficients Cpe for walls of circular-plan buildings" or AS/NZS 1170.2 chapter C5.2.

Of course for engineering point of view is an advantage to work with an unique (equivalent) force coefficient related to the projection of tank area, instead to consider the variable pressure. This is the force coefficient or drag coefficient. Because it is obtained as a mean value by an "integration", it depends on Reynolds number, surface shape but not to "point location".
For me this coefficient is also the "shape factor".

To detail a little: the drag force can be found by integrating the pressure and wall shear stress. Well, the last term seems to be lost in some works dedicated to tanks, for example AS/NZS 1170.2 says "the drag force coefficient arises from an integration of the along wind component of the normal pressure) and gives a value of 0.63.

This result seems to be confirmed by other works/ books where - for example- you can see for Re=10^5 and h/D=1 (cylinder) Cd is 0.63, for h/D=2 we have Cd=0.68 and Cd increases up to 1.20 for an "infinite length", etc.
So I would consider that a cylindrical "shape factor" of 0.6 or 0.7 makes sense for tanks.

By the other hand, ASCE 7 gives directly the "force coefficient" as 0.5 for h/D=1 ratio of 1 and D*sqrt(qz)> 2.5. Again, for me this is also the "shape factor" in this case- and I assume the value is based on more accurate data.

For this reason I think API 650 footnote (b) of 5.9.7.1 is not clear, because they said that due to buckling of shell is no need to consider a "shape factor". I guess this footnote was just a way to escape long discussion; as I tried to explain above, API 650 considers a force factor of value of 0.58 when imposes an equivalent pressure value of 18 psf on the projection of tank's shell vs. 31 pfs as design wind pressure.

Hello Fedeghi,

I would add some suggestion for you.

First, it's about considering a figure on the uplift i.e. suction pressure on roof.
APi 650 considers uplift [30 psf]*[V/120]^2 or 1.44 kPa*(V/190)^2- a value very closed to that one considered in 5.9.7.1 i.e. (31 psf)*(v/120)^2 or 1.48 kPa*(v/190)^2.
So you may chose the value of upper shell elevation wind as roof uplift; alternatively you may observe (and try to convince your Client) that API considers as roof uplift the wind pressure@10m applying a correction by 30/31 to it.

Second, it's about the buckling- Intermediate Wind Girders necessity.
In fact, what API 650 footnotes (b) and (c) of 5.9.7.1 try to say is buckling must be checked considering an uniform external pressure and that pressure is 36 psf*(v/120)^2 or (1.72kPa)*(v/190)^2; that pressure mustn't be corrected by a "shape factor" since this last one is specific to calculate the wind force, not buckling.

So working with "wind profile" you still need to choose a pressure to check shell buckling, i.e. a "wind bucking pressure".
I think you would consider the top shell wind pressure+5psf (or 0.24 kPa) as "wind buckling pressure" and to correct the results as 5.9.7.1 point d. asks for. However, strictly speaking, API operates with wind pressure@10m@V=120(or 190 km/h)+5psf (or 0.24 kPa), the sum being corrected by (v/120)^2 or in SI (v/190)^2.

Just in brackets, the modified U.S. Model Basin formula that API mentioned in footnote c is of form
p_cr=[2.42E/[(1-m^2)^0.75)*(L/D)]]*(t/D)^2.5
E is elastic modulus, m is Poisson coef, etc

This formula is the basis of 5.9.7.1 where p_cr is considered the wind pressure chosen for buckling calculation. So that formula already includes 36 psf (1.72 kPa) and considers numerical values for E and m.
The same formula is the basis of "transposing width procedure" based on the top shell thickness, which means that a shell course of width W1 and thickness t1 has the same p_cr as a shell course of width W2 and thickness t2 when (t1^2.5)/W1=(t2^2.5)/W2 and this leads to 5.9.7.2 a.

It is funny that based on your variable wind on shell and "US Model Basin" formula, the things just got more complications because the critical pressure is decreasing course by course...

Sorry, I mistyped "pressure@10m", instead I think is "pressure@40ft" or 12.2 m

Best regards.