Hi !!!

For Flange leackage we r using equi pressure method and if it fails in that method then we r going for the CAESAR calculation method.

CAESAR calculates Operating and Seating stresses and it compares with the allwoables.

Can anybody tell me, which formulas r used to calculate those stresses specially for seating stresses ?

Now my another dbt is, if flange is failing in seating stresses wh to do ?

My few friends suggested me following options :

1. Reduce the allowable stresses of Bolt material and chcke its seating stresses, still it is failing then
2. Make Pressure equal to zero and check flange for external forces and moments, still is failing then
3. Check it with actual design pressure, force & Make moments values zero, still it fails then
4. WRITE NOTE : " since flange leackage is in seating case it can be neglected."
Pleaze can anybdy tell me the neccessity of checking flange in seating stresses ?

and if it is failing then what assumption to be taken and wht will be its base ?

Flange stresses are calculated in accordance with ASME VIII div 1. The method is same in PD5500. ASME VIII div 2 gives different equations but similar results to div 1. Div 2 includes external loads directly without need for the 'Kellogg' equivalent pressure method.

It is common to find that certain B16.5 flanges are over-stressed in the seating case but ok operating. 150# to 400# are the worst, particularly when a spiral wound agasket is used. This is an incompatability between B16.5 and ASME VIII. Clearly there are many thousands of these joints successfully in operation so it is not a real problem.

If you are dealing with low pressure flanges, you might consider the CEN EN 1591 code.

Hi vinay,

"Gasket seating is the initial applied bolt load for assembling the joint at atmospheric temp and pressure. this is a function of gasket material and surface area occupied by gasket"

Please read ASME sec VII div-1,appendix-2 for detailed flange design. In para 2-7 you will get the formula for calculation of flange stresses.there are three basic formula available for:
1. longitudinal hub stress
2. Radial flange stress, and
3. Tangential flange stress

Please note that above formulas are common for calculation of stresses in operating case and gasket seating case. only the difference in these formulas are the value of " Total moment acting upon the flange, for the operating condition or gasket seating condition"

Allowable stresses for gasket seating and operating condition is the allowable stresses value of falnge material at atmospheric temp and design temp respectively which you can get it from B31.3 appendix A, table A-1.

In operating condition, the total flange moment is the sum of three individuAL moments in same direction mentioned above for stress.

For gasket seating, the total flange moment is based on the flange design bolt load, which is opposed only by gasket load.

Hope this will work for you.

Sorry typing mistake..
its ASME sec VIII div-1,appendix-2

1. Raju Soni has admirably pointed out the case. True, the difference in the formulas (operating vs. seating) must be only the value of "Total moment acting upon the flange, for the operating condition or gasket seating condition". And yes, for gasket seating, the total flange moment is based on the flange design bolt load, which is opposed only by gasket load.

2. It would be the moment for seating condition greater than the moment for operating condition, leading to a J index over 1? I made some numerical tests for 150# flange cases. The moment acting upon the flange for seating condition was less than the moment for operating condition….so for certain B16.5 flanges I would say they are not over-stressed in the seating case and ok for both seating and operating, despite the Caesar calculator predictions! Moreover, looking to the mathematical relations there is a big chance that this remark would be true for a wide class of flanges 150# to 400# with spiral wound gasket, so I cannot confirm there is an incompatibility between B16.5 and ASME VIII.

3. In fact how does Caesar calculate the moment for seating condition?
In my opinion, it would be simple calculated taking into consideration exactly what Raju Soni highlighted: considering the reaction on gasket for seating condition that should be exactly W0 - the flange design bolt load.

Can anyone give me an example of a flange size & rating that exhibits this incompatibility between SecVIIIDiv1App2 and the equivalent pressure method?

Vinay,

Sec VIII DIV 1 APPENDIX 2 , APPENDIX S AND APPENDIX Y are not the only sources for doing a flange analysis. SEC VIII DIV 2 , SEC III and B31.8 also has methods. For a comprehensive view of the different applicable methods see Compansion Guide to ASME Boiler and PV Code ( edited by K.R.Rao) chpter on Flange analysis.There are other standards also like say EN1591 but it is too involved.

Regards

Steven,

The incompatability I mentioned is not to do with the 'equivalent pressure method' which is used to include external loads into a flange analysis otherwise based on pressure alone.

My reference was to many ASME B16.5 flanges in the lower pressure catagories, when analysed using either ASME VIII Div 1 App. 2 or ASME VIII div 2 section 4.16 (2007), and incorporating a hard, wide gasket such as a spiral wound type, then apllying the ASME B16.5 allowable pressure.

These flanges will often fail the seating stress limit but be acceptable in the operating case. This makes no sense. In addition, countless such units have been in use worldwide for decades.

I try to reformulate Steven’s question.

Can anyone give us an example of an ASME B16.5 flange, size & rating, that exhibits "the failure" for seating stress when analyzed using either ASME VIII Div 1 App. 2 or ASME VIII div 2 section 4.16 (2007)?

Before to send it, please quick check that the total flange moment is based on the flange design bolt which is opposed only by the gasket load, or alternatively, please send some details on the gasket load that has been considered in the calculation.

Thank you.

Try 3", 8", and 14" Class 150 flanges. They are the usual suspects. You could have figured this out by yourself if you had just plotted total bolt cross-section area vs. flow area for the Class 150 flanges.

It is an insult to the members of this board to post and re-post a question without, apparently, doing any thinking about the problem on your own.

Dear board of this forum,

I’m happy to inform you that my company has bought Caesar II licenses and one of them is for my use under the contract with my company.

This statement shows few benefits which I try to list bellow, in the importance order:
- CraigB must postpone spitting on me (CraigB, I’m not "one of those scum" so please, don’t spit on me….);
- I have some legal rights and responsibilities and I understand them;
- Particularly I have the right and the privilege to put questions (and I try- in the limits of my poor knowledge- to not question stupidities, but I cannot give you any warranty on this matter…;
- Particularly I have the responsibility to clarify the Caesar II results that seem to be not complying with the engineering practice. At least, in the "Disclaimer- Caesar II" section it is clear said "IT IS THE USER RESPONSIBILITY TO VERIFY THE RESULTS OF THE PROGRAM" and I take it seriously.

Please consider my questions as a part of my effort to "verify the results of the program". Please note I’ve tried – in the limits of my understanding- to clarify myself some aspects and I failed. I failed also to find out some particular clarifications in the Caesar II papers.

So please don’t get so easily offended on my insistence to question or double- question.



Regarding the particular case of the “Flange calculation" I have some comments:

1. Mr. Richard Ay has answered in the “Flange analysis in Equivalent Pressure method" topic

Quote
The Hg force is derived from the equation:

2 * b * Pi * G * m * P for the operating condition
Code:
b = gasket effective width
G = gasket effective diameter
m = A pressure multiplying factor
P = The pressure acting on the flange
Quote


I find this approach in contradiction with one American fundamental work that is the Taylor Forge’s "Modern Flange Design Bulletin 502".

Since “IT IS THE USER RESPONSIBILITY TO VERIFY THE RESULTS OF THE PROGRAM", I’m very interested to know if Coade is going to change this approach.

2. Regarding the seating condition– which is the subject of this topic.
Based on this criterion, the Caesar II calculation often fails to qualify a lot of flanges and I can see we prefer to say “it is an incompatibility between B16.5 and ASME VIII". I said I made some numerical tests and I cannot confirm this conclusion.
Since “IT IS THE USER RESPONSIBILITY TO VERIFY THE RESULTS OF THE PROGRAM", I’m very interested to know what particular Hg – gasket load- Caesar II takes into consideration for the seating case?.
Please note that I cannot find out this clarification in the Caesar II documents, so my question is not trivial.

With the hope my remarks don’t offend the honorable board of this forum,

Thank you in advance and please consider my best regards.




Dear CraigB,


If you pay the effort to read my precedent posts you can see I said “I've made some numerical tests for 150# flange cases." So really I don’t understand your accusations.


Regarding the "seating condition check".
The matter is why a particular flange is not passing the seating condition.
For the seating check, I cannot see a connection “total bolt cross-section area vs. flow area".
Moreover, the weaker the bolts are, the less the seating condition stress is.
That not means, of course, this is the problem’s solution….the bolts load must assure “Y" stress on the gasket.

Best regards,

Mariog,

I was trying not to involve in this discussion, however I could not stop myself. I would like to give the discussion address here for the same subject:

http://www.eng-tips.com/viewthread.cfm?qid=204661

This discussion is for the one of the flange types that GraigB was adressing above.

I believe you already have a spreadsheet since I follow your discussions on this forum on the flange connections. Therefore you must understand that there are some flange geometries with small number of bolts, bolt sizes and large flow area. Second is the use of spiral wound gasket which requires large seating force.

If you put them together you will find these gaskets are failing under the seating condition or/and the bolts are failing. It is nothing to do with the calculation method. CAESAR II follows what codes. I have my own spreadsheet and I see these failures even without consulting to the CAESAR II flange analysis.

I case there are external loads the problem gets worse.

The thing I do not understand is how ASME B16.5 accept these flanges with spiral wound gaskets.

I guess that you are trying to understand how Coade solve the leake calculation. I believe this was discussed again many times. As far as understand they are using the stiffness of bolts, flange and the gasket in their calculation and check against the loading.

Hope it satisfies all parties. Kind regards,

Ibrahim Demir

Dear Ibrahim,

You can see I’ve mentioned a mistake in the Caesar calculation, in the equation used for the gasket load. I think it is very constructive to understand and to correct the possible mistakes, and that’s why I’m very confident Coade is proceeding to correct this mistake.

Let’s do a step-by –step approach and first to understand what is the gasket load that Caesar takes into consideration for the seating gasket load. This is information you cannot find in the Caesar documents. Neither the gasket load equation under the operational condition can be found within these docs.

Thank you
My best regards

See the revised (earlier) post. I referenced Hp instead of HG.

http://www.coade.com/ubbthreads/ubbthreads.php?ubb=showflat&Number=18926&page=1&fpart=2

Dear Mr. Richard Ay,

Thank you for your answer.

An error in a post is fully understandable and practically inoffensive.
Unfortunately, I still claim there is an error in the calculation.

The "gasket seating" results and "operating condition" results must be practically the same for a very small pressure value as load.

Please try with a small pressure value, let's say 0.1 psig as a pressure load (and by the way, this wouldn’t be referred as "design pressure" load, it is an "operating pressure" load- in my opinion, the terminology is quite confusing).
The discrepancy (between the "operating" and "seating" results) cannot be real.

With the hope you’ll correct also the software,
My best regards.


PS. I just suggest you to carefully evaluate the equations. As I said in that thread, the gasket load at p_max is HG=2 * b * Pi * G * m * P_max , but this is not an equation for all "p".

CAESAR II (and PVElite and CodeCalc) implement the rules of ASME Section VIII Division 1 Appendix 2 for the flange computations. It is standard industry practice to use "Peq" instead of "P" in the appropriate Code equations.



Why?

To help explain the results please download the ZIP archive here. In this archive is a PDF of the Taylor Forge (Example 1) document, this example as run both CodeCalc and CAESAR II. (The CodeCalc document includes equations with substitutions so you can see exactly where all of the values come from.)

Also included in this archive is a second job (TayorForge2) where the pressure has been reduced from 400 psi to 1 psi. Note the seating bolt stress does not change, while the operating bolt stress does.

Richard,

I did not see any problem with the Coade flange calculation which is based on the codes. I guess the pipe stress engineers want to see the results with the formula used in the output similar to PVELITE printout that you have provided. Sometimes, the output result only become meaningless if you are net very familiar with the codes or if you are trying to explain how to solve problem in case there is a failure. Perhaps the step-by-step calculation may give stress engineer a bit more comfort on understanding what they are doing.

If you can provide the similar output in the CAESAR II Flange calculation, I guess, this kind of discussions will end.

Kind regards,

Ibrahim Demir

I'm not going to do that. Over 60% of the source code in our vessel products deals with "equations with substitutions". It is a tremendous source of problems and headaches. Not to mention the fact that if you do anything novel or unique you give it away to the world.

Richard,

I suggest to make this topic one of the sticky topics on the forum to cut this kind of discussions on the subject. So, the topic reader can find a sample problem to compare with code rules. I believe that most of the new joiners will not go through the search option.

Kind regards,

Ibrahim Demir

Dear Mr. Richard Ay,
Thank you for the partial answers.

1. The "gasket seating" results and "operating condition" results must be practically the same for a very small pressure value as load.
Why?

Because in terms of mathematics, the forces, stress etc must be continuous functions against the variable "p". A discontinuity is possible but must show something very special.

In terms of physics. In your example for 1 psig load, the forces reported (in Plot) are HG=797 N, W=1371115 N.


Since for seating condition HG=W, what exactly is the physical phenomenon that is able to modify the HG to 0.058% of the initial value? For a very small pressure load, the gasket load must have a value near the bolts load, that’s my opinion.

2 I can see an axial force is not affecting the reported HG. Please, can you give some details: what reference is following this approach? To be clearer, the Kellogg results, very common referred, show exactly how the gasket load is affected by the external loads.
If I understand well the results, a pressure load is changing the gasket load while an external load no?

Best regards.

I agree to your terms and conditions. Welcome to the legit world of pipe stress analysis, Mario.

An agreement is always a progress.
As I said in another post, I really appreciate the knowledge you are sharing, it’s a high quality one.
Just know the word "empathy" is quite specific to English, so it is suppose to be more accessible to you. In a lot of languages there isn’t a direct correspondent word; as likely as not we’ll act based on presumptions.

Best regards,

…while I can see a problem in!

The ASME VIII equations are specifically written for the design case.

Exactly this is the problem: ASME VIII includes equations for the design case.
That’s why the Flange calculator is an excellent design tool. And that’s why for the purpose of the piping flange checking under the operational loads (operational pressure, operational axial force and operational bending moment), the flange calculator fails until the point the evaluated forces are not in mechanical equilibrium!


For the piping flange checking case:
-It would be more realistic a W sensitive to the bolt initial tightening stress, but we haven’t such equation in Code.
-The Taylor Forge "operating" HG=W_m1 – H isn’t a Code equation. If you are working with substitutions in this equation, based on the relation(1) of the VIII Div1, appendix 2, the result would be HG proportional with pressure, which is not true.
More realistic is the function HG= W-H= W- P* Pi/4 *G^2 for the operating pressure load. Or, if we have p, F and M loads, more realistic formulas are
HG= W- P* Pi/4 *G^2- F- 4*M/G or
HG = W- P* Pi/4 *G^2- F- 4*M* [I/(0.3846*Ip+I)]/G if you accept the Koves theory. But these formulas are not Code formulas.

The improved Flange Calculator would be based on such equations that are not really ASME Code equations, but are in agreement with the ASME approach!

PS.

Ibrahim,

still want "to make this topic one of the sticky topics on the forum to cut this kind of discussions on the subject"? :-)

My best regards,

It is worth to mention that it would be more appropriate to have a "stress-strain" analysis of the flange joint.

A "Stress-Strain" analysis is not banned by ASME VIII, since Div 1 describes it in Appendix S: "Such analysis is one that considers the changes in bolt elongation, flange deflection and gasket load that take places with the application of internal pressure, starting from the prestressed condition".

Or in an ASME book, under the "Recommended approach" title: "The most efficient approach for evaluating leakage is to evaluate the flange as a structural system. The flexibility of the flange gaskets and bolts must be modeled to accurately predict the gasket and bolt loads resulting from pressure and external loads”.

Not really interested to have such approach?

Best regards,

Have a look at EN 1591 the method does exactly what you are proposing, a stress - strain alaysis of the flange, bolts and gasket assembly. It needs alot of data, but so would any accurate appraisal of flexibility and interaction in such a complex shape.

True. Is Coade interested in implementing it?
Would an USA Client to accept such calculator based on a European Code? Middle East? Every European company?

Best regards,

We can take a look at it.

Great! and thanks! But this would be a long-term project for Coade.
About the subject, what is your intention in the near future?

Regards,

CAESAR II computes flange stresses according to the ASME Code equations. Until ASME revises these equations, or some other Code publishes alternate equations, there isn't a whole lot we can do.

I understand your position.
Yes, it is a strange story.

By one hand, if we try to change something, we haven’t a validated ASME approach and we would only guess something congruent with.
By the other hand, maintaining "as is" this ASME design mathematical model for the flange checking purpose, we have strange results, which is evident when the pressure loads are substantially less than the rating pressure.
Including the external loads is possible only via "PEQ" methods just because some PEQ formulas are included in ASME Codes or worldwide accepted. But the truth is there isn’t a PEQ formula able to make a true equivalence, because the flange stress under a bending moment isn’t axial- symmetric. One formula is able to count an equivalent stress on gasket, but the flange stress is not proved as being equivalent. And so on…

Best regards,

Dear MoverZ, Dear CraigB,

I don’t know if I can convince this forum there is a serious problem in the Flange Calculator. Anyway this is what I’m thinking about and this is my (legit User!) feedback to Coade.
The Calculator applies the flange design equations as valid for every pressure load because CAESAR II computes flange stresses according to the ASME Code equations.
The only question would be if we really understand the ASME Code equations.

About the particular aspect of the bolts, we all know the ASME Code counts a "seating" bolts load and an "operational" bolts load. Caesar II Flange Calculator evaluates both.

I’d like to convince YOU that the ASME Code doesn’t really consider a variation in bolts load!
Please don’t say yet "you are a ……!" (even this is possible!), and PLEASE take a look to the pages 17- 19 of the Taylor Forge 502 Bulletin.
I’ve attached these pages.
You may observe that the figures 4, 5 and 6 of this reference show that, when a pressure is applied:
- the bolts load remains as W
and
- the flange is always under mechanical equilibrium.

The Taylor Forge’s work is assimilated by the ASME VIII Code, so for me, it makes sense to put the question "Really we understand what the ASME Code says?"

Let me explain my understanding.

The ASME Code (Rossheim- Marks- Taylor Forge) wants to assure for gasket two "magic" numbers "y" and "m":
- For the seating condition, at least the value "y";
- For the maximum pressure load (let’s say "design pressure", p_max) at least the value "m*p_max".
The bolts must assure the most conservative condition.
For this purpose, the ASME Code procedure is to calculate two required forces and two required bolts areas and to develop the calculation selecting the biggest value.
W_m2 is the bolts required force able to assure the value "y" on gasket, for the seating condition.
W_m1 is the bolts required force able to assure the value "m*p_max" on the gasket, for the p_max load.
The ASME Code equations lead to a value W that is greater than W_m1 and W_m2, so there is the certitude that the actual values "y" and "m" shall be greater than those tabulated.
The bolts are selected.

And that’s all about the bolts; here the Code finishes his intentions.

Did the ASME Code say that the actual bolts load is W for the seating condition and W_m1 for the operational condition? In my understanding, no.

If I want to understand more about "operational" loads, the Code equations are not really helping me.

In fact, when tightening the bolts, the bolts load is not W_m1, is not W_m2 and probably is not W. The actual value is done by the applied torque moment and consequently by the bolt tightening stress. And I can note this value as W0.

I may assume that W0 remains constant.
Is this real? Probably not, but this assumption has been conservatively taken for dimensioning case. This was the Taylor-Forge approach, their equation are ASME Code now.
If I want to know more, I need a "stress-strain" analysis of the flange and I’ll understand how the bolts are participating.
If I want to repeat the same assumption that has been made for flange dimensioning, I would consider W0 as constant.

I can now verify the flange. There is no need to check the gasket stress but I can do it.

For the seating condition: the gasket load and bolts load are under mechanical equilibrium. So the gasket load is W0 and the actual "y" value is greater than the value tabulated. I can calculate the flange stress, J index, etc. for the seating condition.

For any operational condition (a pressure load p<p_max), the gasket load, bolts load and pressure load give forces that are under mechanical equilibrium.
I can evaluate the gasket load as W0- Pi/4*G*G*p. This assures the mechanical equilibrium. The actual "m" value would be calculated and is greater than the value tabulated. I can calculate the flange stress, J index, etc. for the operational condition. Because the assumption made, the bolts load remains as W0, so the operational bolts load is assumed to be the same as seating bolts load.

What do you think about my interpretation?

My best regards,

NOTE. I've corrected few typing mistakes in text. Sorry!



Dear MoverZ,

You say "Div 2 includes external loads directly without need for the 'Kellogg' equivalent pressure method.".
Not exactly...I would say.
For the design case, it is perfectly true. For the operating case, another story....
The Code is not giving us an equation for the gasket load under the external loads, so how we are going to count this? I’ve propose two formulas, following the same interpretation… but, in my opinion, they are too conservative! Again, a "Stress- Strain" model would say much more about.


Best regards,

Good grief. How many thousands of flanged joints have been sucessfully designed by the Taylor Forge method ? Conservative it may be but it most certainly works, albeit with obvious technical limitations.

Maybe you could stop bleating and use EN 1591 or 3D FE analysis for your flange design.

MoverZ,
Dear Good Shepherd,

I’m bleating on piping flange checking. For the flange design I have no bleat. :-)

About EN 1591. The Mechanical Model details are saying "Only circumferential stresses and strains in the ring are considered. Radial and axial stresses and strains are neglected".
Applying the model for ASME Flanges could be a reason for some bleating?

3D FE for piping flange checking just sounds great, as engineering! Some additional project cost...another bleating!

My best regards,

All ASME flanges.
Basically, the ASME Method is the Taylor- Forge method.

best regards,

Some details on EN 1591-1 method.
This is an Advanced Solution Engineering Ltd document.

Best regards,

Mariog,

I am well aware that the Taylor Forge method is the ASME method of flange design. It is also found in practically identical form in PD5500 and other international codes.

And…the same limitation in all….
The "Kellogg" PEQ method is based exactly on the same assumption: the bolts are not participating; the external loads produce effects only on gasket. But here, because the bending moment, the gasket stress is not constant. The maximum gasket stress is considered equivalent with one produced by an additional pressure.
Yes, it certainly works, albeit with obvious technical limitations…


best regards,