Hello board.
I have a "head scratching" problem/issue.

Sorry if it is long but you need the facts to advise on it.

I design a small vessel for client to operate at 60bar/6Mpa with a 200NB shell and a ring flange and blind flange to seal off the top end.

The bottom end has a "dished" end. The material of the body is SA240 304. The ring flange has an "O" ring groove to seal against the blind flange. There are 8 holes in the ring flange with 1" SA193-B8 studs and nuts.

No corrosion allowance considered and the "rules" say the thickness of the flange must be taken at the root of the "O" ring groove.

I started with a 25mm thick ring flange for the calculations. The groove is 4.5mm deep so the thickness for the calculations is 25-4.5=20.5mm.

The problem is the design fails for the ring flange due to tangential bending stress but the blind flange is (almost) OK. The design suggested 29mm approx. for the blind flange, but 45mm for the ring flange.

This is confusing to me because the flanges react to similar forces.
(Note: ***NB*** with O-ring type of seal, there is very little resistance to bolting (seating) so the "m" and "y" gasket factors effectively = "0". This makes the calculations more simple to do.)

For those "in the know", the H force (hydrostatic end force) acting on both flanges is

(PI.G^2/4)xPressure ,

then this must be multiply by moment arm "hd" distance and consider the formula for stress is

Y.(Bending Moment)/(t^2.B) , where "Y" is a factor and "B" is flange inside Diameter.
Note, the "t" is at the root of the gasket "O" ring groove so is 20.5mm

This gives a very high stress value for the ring flange of 356Mpa which is much more than allowed at 25deg operating temp. of 130Mpa allowable.

OK, so this is the result for the ring flange containing the gasket.

But it is a surprise to me that the blind flange calculation gives much lower stress. The formula for checking the thickness of blind flange is

G*SQRT(0.3*P/S) where "S" is allowable stress and "G" is gasket reaction dimension which for the "O" ring I just assume it to be the diameter of the groove where the ring fits into.
There are some other parts to this equation but they become "0" due to the "m" and "y" factors being considered "0" also.

The thickness is (almost) OK, the stress is just a little above allowable.

So can someone explain why so different results but the pressure on both blind and ring flange is same? Or maybe I am not understanding the stresses involved.

Thank you for your help.

Dear mgeorgiadis,

1) you are considering making both the flanges from plate material.

2) During the calculation of diameter at location of gasket load reaction, you have obtained effective gasket width as W/8. (table 2-5.2) (where W = o ring diameter)

3) For the operating conditions, the total flange moment Mo is the sum of the three individual moments MD, MT , and MG : in your case MG will be zero, there will be only MT and MD.

4) For the gasket seating condition load acting on flange will be as per 2-5 (e) (5) and moment will be 2-6 (6)

5) calculation of blind flange thickness shall be as per UG-34 (c) (2) and also pls refer fig.UG-34 (K) . Pls refer the explanation given for sketches j and k.

Hope this helps.
Regards
ASD23789

Dear "ASD23789"
Thanks for the summary. The information you give shows me the calculation is done properly because I follow this same formulas you give. By the way, my MD is also zero because there is not a hub.
So then the question remains, why has the blind flange a much thicker requirement? Does it not react to the same forces and bending moments?

As per original post the thickenss of blind flange is G *sqrt(0.3*P/S), the second term realted to bolt moment in equation (2) of UG-34 (C) (2) is ignored (as per my understanding from the original posting).

Apart from the reply given by our friend, I would like to add that theory of circular plate under pressure is different from annular plate. You may refer to any Timoshenko book on Plates and Shells or any other book.

As per Mr Mgeorgiadis, the pressure is same and hence stresses should be same. It may be noted that even for annular plates, stress varies if a/b varies (Y varies in flange formula). Your argument is hydrostatic end force coming within ID of flange is acting in addition to pressure acting on flange per say. Yes it is true that total pressure force acting on ring flange / blind flange or same but the results will be different as per their behaviour

Hope I have made my views clear

With regards

_________________
R.SRINIVASAN
Sr. VICE PRESIDENT
Head, Design and Engineering
PED,ISGEC,
YAMUNANAGAR
HARYANA -135001
INDIA