Appendix 2 Flange design (again) 1

[mgp]

When calculating flanges according to ASME VIII, div 1, appendix 2, in some cases the bolting will fail even when using the tabulated pressure ratings of ASME B16.5. The allowable bolt stresses have been obtained from ASME B31.3 as defined in section 304.5.1.

As an example a 12” 2500# RTJ flange with A193 B7 bolts (specified suitable for all flanges) will fail even with the softest ring gasket (soft iron). There are no stronger bolts (for bolt sizes above 2,5”) in ASME B31.3 as far as I can see.

I am aware that of the bolt considerations for bolt stresses of Appendix S of ASMEVIII, div1 where it is explained that the allowable bolt stress used for the calculation will be exceeded in some situations, but it does not say that you may initially select bolting with available bolt area less than that required by appendix 2.

Is anybody able to explain how this is possible or what I may have overlooked.

I made a flange design spreadsheet which I have sent to lots of you members thread378-19042, and the question came up from one of the recipients. I have made a cross check with another calculation programme (not by myself), and it came out with the same result.

This question is somewhat related to a previous thread794-51573, but in this case it does not seem possible to make an acceptable joint by changing bolts or gasket so by doing a calculation on an (according to B16.5) acceptable joint, the joint fails.

For reference I have attached all data I have used and the formulas used. No flange material data are included since these don’t affect the formulas of this question,

Here goes:

Base data:
******************************************
12” Flange ASME B16.5 Class 2500
Design Pressure 6170 psig at 100 F

Gasket – Octagonal Ring Joint - soft Iron
From ASME B16.5:
m=5.5,
Y=18.000psi
Gasket diameter = 16.0”
Gasket contact width = 1.25”

Bolting
12 no’s 2 3/4" stud bolts
Bolt root area per ASME B1.1 : 5.26 sq-in (section at minor diameter)

Materials
Bolting A193 B7 (2.5”<d<4”) (Acc B16.5 section 5.3.1 may be used for any flanged joint)
Allowable stresses per ASME B31.3 : 23.0ksi

Calculations:
********************************************

Basic gasket seating width
b0 = w / 8 (RTJ) =0.15625 in

Effective gasket seating width
b=b0 for b0 < 1/4&quot; = 0.15625 in

Gasket load reaction diameter
G = mean gasket diameter for b0 < 1/4&quot; = 16.000 in

Total joint-contact surface compression load
Hp = 2 * b * pi * G * m * Pd = 533,050 lb

Total hydrostatic end force
H = pi / 4 * G^2 * Pd =1,240,552 lb

Min. required bolting load for operating conditions
Wm1 = Hp+H =1,773,602 lb

Total required cross-sectional area of bolts
Am = greater of Wm1/Sb or Wm2/Sa = 76.12 in² (Wm2/Sa will be smaller in this case)

Total available cross-sectional area of bolts
Abolts =12*5.26=63.12 in²

CONCLUSION: required bolt area is smaller than required => the bolts will fail.


Regards
Mogens
mgp@mxl.dk

 

[prex]

Agree with your calculation, though the required area is even a bit higher (77.1 sq in).
However can't right see your point: the depicted situation is well known (at least to me), but, if you are under a regulation that allows for the use of B16.5 ratings without recalculation, then you are OK, otherwise you'll need to derate the flange.
If you worry about the possibility of bolt failure, don't forget all the margins available, with a supplemental one for bolts: the root area that is used in flange calculations is a safe parameter for determining bolt resistance, a more realistic figure would be somewhere midway of root area and area corresponding to outer diameter.
The calculation of your flange may be seen on the site below (till someone changes the data), under Vessels -> ASME VIII-Div.1 -> App.2 -> 2.7:int.

prex

http://www.xcalcs.com

Online tools for structural design

 

[Elvie]

mgp,
although your calculations are about to be correct (see prex adjustement )the design case you have selected is quite unrealistic.It is not usual to design P.Vessels/piping at full ASME 16.5 material group rating in 1500# (or even 900#) and 2500# pressure classes.
However in the case these design conditions are really applying you may have to choose between the following alternatives:
1) use hub connectors instead flanges,or
2) use compact flanges,or
3) if you have to stay on ASME B16.5 flanges,do not bather;tension (not torquing) your bolts at +/- 50% of their SMYS i.e. 50ksi(which is well above the 28099psi needed to satisfy the design conditions) and the problem is solved .In this case you are going to sustain the hydrostatic pressure testing and then have enough margins to face all operating conditions for a long periode of time.

 

[mgp]

Thanks Prex and Elvie for the reply.

Now at least I know that my calculation was correct. (The 77.1/76.12 sq.in. was a copy-paste error – I used 23.0 ksi for bolt stress)

However this leads me to another question (sorry if it is a bit long):

I must admit that I was not aware that this was a well known thing, or that it was unusual to use full class ratings for 900# to 2500# systems.

Most of the piping classes I have seen (including high ratings) have service limits equivalent to that of the corresponding ASME B16.5 rating tables. One example can be seen here:

http://www.standard.no/standard/NORSOK_standards/L-001/pclass/GC11.htm

This is one of many piping specs specified for use throughout the Norwegian Offshore Sector (North Sea Europe). I’m not saying their specs are correct – just a bit frightening if they are not.

In individual piping systems, the design pressures will off course normally be lower, but sometimes it comes close and this is where the flange analysis comes in:

-Pipe stress calculation is made
-External loads are converted to equivalent pressure
-The total pressure is compared to the piping class (B16.5 rating)
-If it fails – an Appendix 2 flange calculation is made and ----ooops---- the flange failed -even before the external loads were applied.

According to Jsquare in thread794-51573 flange checks made by CEASAR II doesn’t even check the rating but perform the calculation right away.

We know that the bolts are really ok, because of the conservative design loads, but we cannot produce any authorized evidence (read calculation) to prove it.

To me this leads to the following recommendations:

Piping system Flange check:
The sum of design pressure and equivalent pressure due to external loads must not exceed the listed ratings of ASME B16.5 or the rating calculated from ASME VIII, Appendix 2, whichever is HIGHER (Not lower)

Piping Class Design
Service ratings for piping classes using ASME B16.5 flanges should be set to not higher than B16.5 listings, or ratings calculated by appendix 2, whichever is LOWER. (This will ensure that stress analysis will pass at least with no loads)
(Elvie - is there any other recommended margin?)

Please comment

Regards
Mogens
mgp@mxl.dk

(Please use this mail address for the spreadsheet - the old one is out of date)

 

[LSThill]

Team Members

ASME B31.3 304 PRESSURE DESIGN OF COMPONENTS:
Piping Design Classification has two cases:

1 For NPS 3/4 through NPS 48 (Full flange ratings per ASME B16.5, Table 2-1.1.)

2. For NPS 3/4 through NPS 48 (Pressure and Temperature Rated Spec. Limited by Pipe and Full Vacuum).

Must check your calculation for component the above two cases.

lst

 

[mgp]

Isthill

I don't exactly see what you mean.

Piping specs shall off course be created to satisfy both of your points 1) and 2), but the point is even then, the flange fails if one is forced to do an Appendix 2 analysis.

regards
Mogens

 

[Elvie]

mpg,
first of all you may see also my point in Thread292-60168.I may have to enhance here my general oposition in using the concept of &quot;equivalent pressure &quot;under piping code B31.3.
For HC plants the flange leakage may be of a big concern there where huge HC inventory may be released.This may be related to P.Vessels only.For a 2500# system I hardly see an application except Test Separators on HC Gas fields generally designed at WHSIP( well heads shut-in pressure).But even here I cannot see using 12&quot; test headers.However I have to admitt we may have to deal to an exceptional case where the Test Sep is fully 2500# rated and have also piping connectios with 12&quot;X2500#ASME 16.5 flanges.If this is the case an interactive piping stress analysis with the piping routing and supporting must result in very low if any external piping loading on the flange and nozzle.If you still intend to use the equivalent pressure concept then the total load must be compared with Wm1 (the required bolting load for op. conditions as you have calculated)as the equivalent loading is sure will exceed the ASME 16.5 rating (design pressure being the material group pressure class rating).
Easier will be to calculate the actual bolt stresses at maximum piping loading and tension consequently the bolts (eventhough you exceed the bolts allowable stresses in B31.3)with a SF of two but not to exceed 50-60% of the bolts SMYS.
For flange connections on piping ,again I recommend using hubs eventhough the hubs are more expensive in the first place(C.S. case).But the additional benefits,stronger connections and largelly lower make up time will more than compensate the cost differential.And do not forget in an HC Plant the HC detection system is there to protect.

 

[bvi]

Treating loads on pipe flanges as equivalent pressure is extremely conservative. You cannot assume you are OK because the stresses are within the allowable stress for the pipe. The equivalent pressure approach is highly conservative, and will fail many flanged joints that are actually acceptable.

One of the reasons (and there are several) that it is conservative is that the gross bending moment on a flange is not actually equivalent to an axial load. The torsional rigidity of the flange acts to distribute the bending load around the flange. There is a paper by Dr. Koves on the subject (it may have been published in the Journal of Pressure Vessels and Piping so you may be able to pick it up in a search). It is also included in proposed new Section VIII flange rules. Dr. Koves has also written a computer program, Kflange, which includes consideration of flange torsional rigidity in assessing the bending moment. I use this on occasion, usually when troubleshooting leaking flange problems. Typically, designers don't check the flanges for bending loads, and typically, they are OK. However, this doesn't mean that they are not responsible for designing a system in which the flanges don't leak.

Some people use the equivalent pressure for simply ranking flanges, or for only consider it an issue if the equivalent pressure is some factor (e.g. two times) the pressure rating.

Making sure the bolts are properly torqued, say to 50,000 psi prestress, goes a long way towards making the flanges resistant to leakage resulting from piping loads. The preload in the bolts, beyond the basic allowable stress used in bolt sizing, is what keeps the joint tight when there are piping loads on the flange.

 

[Alengfo]

I havent studied your calcs ... some very competent people already have. What I can share is that I have tried to re-analyse and recalculate many flanges using all sorts of software incl FEA , after many hours/years of manual calcs. Frequently they do not pass , as is, existing and including the Grandfathered in B16.5 standard flanges . All this without any loading, Pe or otherwise.
What I have learned since is that is not uncommon, (what a lot of wasted effort) and that standard flanges do not pass
including under appendix 2 analysis.
There is a lot of heated dispute in this area, esp. around revision of M & Y factors and how the methods work eg.
eperc.jrc.nl/documents/pdf/bulletins/Bulletin_3_PDF/ EpercBulletin_3_contents.pdf