Hello everyone..


We have a model of relief valves at the top of a column where Peq. is twice of the Allowable Pressure, I would like to know if it's possible to decrease de % Ratio without changing the arrangment.

We have noticed in some configurations when we evaluate the Peq method tends to fail when flanges are smaller than 6 in. of diameter due to G value.

Regards.

Dear Wromero,

The solution should be change the arrangement to reduce at least one of axial force or longitudinal moments. Increasing the rating of flange let say from #150 to #300 is seldom done in order to increase allowable pressure for the same size and teperature.

Peq = 2* Allowable pressure...looks strange.

Hello again..

Is it correct to evaluate Peq under the following conditions:

*(OPE) W+T1+P1 (Line operation condition)
*(OPE) W+T2+P2 (Line design condition)


Or if it is not correct, in which condition should I evaluate the line?


Regards.

wromero,

In case you are using the design conditions to evaluate the connection and the design condition is close to the temperature-pressure coverage limit for the material, the flange connection will always fail due to the butterfly effect.

I would use the maximum operating condition for the evaluation for the flange. If it still fails you are correct in increasing the class of the flange.

Additionally, equivalent pressure method mostly used for screening the flanges in the analysis for the conditions that thay are in. So you can select the same type (class and size) flange connection with the highest ratio of Pequivalent/Pallowable to reduce the number of calculation for the same range of the flanges. And then use the force and moment of this connection in the "Flange Analysis" option of Caesar II to check for strength and leak. If this calculation passes, the flange connection is adequate although the equivalent method gives more than 100% for the ratio.

Hope this helps.

Ibrahim Demir

Thanks Samsul and Ibrahim for your replies.

Ibrahim:
We would like to consider your advice but we haven´t defined a maximum operating condition. We just have design and normal operating condition for each line. In this case, how can I define a maximum operating condition for the line?

On the other hand, in your experience have you noticed some relevant differences in the results obtained by the Flange Analysis and the Peq method?

Thanks very much for your help.

wromero,

The maximum operating temperature and pressure should be given to you by the process department and your design shall follow. In case there is no definition for the maximum condition you will use the normal operating conditions in your design. However, if the system goes higher temperature and pressure that the normal condition, no one will know what will happen without any calculation or stress analysis. You need to inform the precess engineer or client for this issue.

Flange analysis and peq method gives the same ratio. This ratio is very useful if you have many same size and class flanges in the analysis. You can screen flanges in the analysis, and do one flange calculation for the flange with the highest ratio for the same size and class. If the flange calculation passes, all the same size and class flanges pass.

Peq is just a comparison value (for this discussion only), however, flange calculation is a calculation for stress on flange, bolt group, and leak of the connection. If flange calculation does not pass you need to change some parameters including flange class.

Hope it helps,

Ibrahim Demir

Hi,

Please note that CEASAR II is not considering any exteranl forces and moment at the flange for analysis.I feel that Peq method would be the ideal one than CEASAR II.For 150#, for particular size , caesar does give wrong result.

regards

Sha.

Yes it does. Where did you get this idea from?

Dear Sir

I agree with you that my assumption was wrong. Since i have done sofar with Peq method for flange leak analysis.And when I worked with ceasar for flange analysis, say that external axial forces and moment is optional. I understood wrongly.Thank you for higlighting this.

Sha..

1. Personally I’ve found the Coade’s flange calculator as working accurate.
2. If one has doubts about, he may use the method described in 2007 ASME VIII Div2.
3. Even I’m not "so" young and I have a deep respect for Kellogg work, I don’t understand completely the attraction of the Peq methods. A consistent progress in this field has been made after the Kellogg work.
4. In this weekend I’ve tried to re-write in terms of Peq the relations given by ASME VIII Div2. It is not an academic debate; it’s just a "positive" manipulation of the relations.
5. I don’t think is a progress doing this. It’s just to help others to understand more about a quite modern approach. A good question is WHY one needs to calculate a more accurate pressure equivalent, instead to calculate directly by using Coade calculator or ASME VIII Div2.
6. This is my "homework" for people than still think the PEQ method is the ultimate truth. As I said, I don’t think so.
7. Being my homework I can consider as original (but as I said, no so valuable comparing to the alternative methods). May be true or not, anyway is not a plagiaristic work. Please consider my ignorance about other papers having the same subject. May be a better paper on this subject, please give me a reference to use in the future.
8. I don’t exclude it may be a wrong typing, since I’ve allocated a limited "personal-time" budget. If is the case (and I hope no)please consider my apologies.

Neither of the two (OPE) cases can fully describe the loading conditions. A new static case, in the form of W+P3+T3+F1 (OCC) will be more indicative of a relief situation.

Notice that the vent stack (or the discharge line after a relief valve can still see considerable temperature (T3) and pressure (P3)--though often marked as “ambient” on the Line List in T1, T2, P1 and P2.

Using the Equivalent Pressure Method (EPM) to qualify the flange design of a relief valve calls for a deeper understanding of the factors in play, beyond a facile exercise of this new feature of C2v5.1. Not to be presumptuous, I would think that before embarking upon external sources for answers, it is prudent to acknowledge first what the B31 Piping Codes have to say about it. Take one reference, the Nonmandatory Appendix II-4.2.3 of B31.1 Piping Code does narrate somewhat about EPM; by and large its “B” part has been generally regarded in doing a leakage stress evaluation.

I second your thoughts.

Just to be clear: what is in my paper is not my method.
It is 2007 ASME VIII Div2 method, re-written.
If calculated correct, the p+p1 total pressure shall produce exactly the same stress and exactly the same J factor and exactly the same flange rotation angle as ASME VIII Div2 predicts for p and Me and Fa loads.
Because a pressure equivalent method is just an equivalent calculus.
My contribution in adding/ neglecting/ changing factors or the basic formulas, etc is ZERO.

In the paper, the pressure equivalent limits are not my limits, are the 2007 ASME VIII Div2 limits.
So I think the real issue is:
- to accept or not the 2007 ASME VIII Div2 method when evaluating the flanges under a B31 Code
- to accept or not the 2007 ASME VIII Div2 limits as applicable also for piping flanges.
What do you think?

My best regards.

1. Not any opinion about using VIII Div2 for flange checking?

2. The Kellogg formula was friendly enough to be accepted without a theoretical proof. At least I haven’t seen it.
But it was conservative enough to get no troubles in real life, and this is the only important aspect.

I’ve just tried to reload the Kellogg work in order to understand it better.
Frankly, it was not so simple. Must free the mind enough to “rediscover” simple things.
I cannot be sure this is the way that Kellogg followed, but I guess so.
I would say the result is a gift for the community.
And sure, for the today technical aspects my paper means nothing.

It may be worthwhile remembering that the 'Kellogg' method for including external loads in a flange design is reproduced in two forms in ASME III, the nuclear piping code so it does have some street cred. So far as the average flanged joint is concerned, there are so many unknowns such as actual residual gasket and bolt load, gasket spring rate and potential pipe misalignment that there is little point getting too technical.

If you want a truly new approach, and one which requires a huge amount of input data, look at EN1591, same as EN13445 App.G. It considers the whole unit as a bunch of springs and pre-defined shapes, then does a mini finite element analysis of the complete joint, employing a multiple step convergance solution. It also considers external loads.

Yes MoverZ,as usual you’ve pointed out quite well.

In fact, I’m in line with your opinion about Kellogg formula, ASME III Nuclear Code, EN1591 and EN13445.
What about VIII Div2? We avoid it just thinking are too many unknowns?

It would be difficult to convince a client your calculation is following B31.3 code with flanges calculation based on EN code, just to follow a modern approach. A clever client will ask you "what about an ASME code"?
And in my understanding, the EN code is specifically counting all the stuff the Kellogg formula ignores, while the quantity of unknowns for what is in field is fair the same.

We are using Kellogg/Nuclear Code formula as a prudent approach, preferring to stay on the safety side. Yes, the Kellogg formula doesn’t count the flange is not a simply circular plate, doesn’t count that often the gasket is not an O-ring, doesn’t make a difference between a tensile and a compressive axial force.
And finally, how much safety is in?
The answer is not important for nuclear field. The safety is important.
Maybe for other branch of activity it would be reasonable to know a safety factor and to make a sound engineering analysis.

Let me make a remark. Suppose you are over the traditional PEQ limit and you are qualifying the flange based on Coade calculator. This is quite the same as qualifying the flange based on J factor of the Section VIII Div2.- this is my opinion.

My best regards

I need to stop myself. What is too much…is too much.

I think there is a reconciliatory way for all the stuff.

1. Psychologically, the first step is to understand what the equivalent pressure is.
For me it is just a mathematical substitution in a mathematical model which is based on a physical model. The mathematical substitution result cannot be better than the physical model.

The Kellogg model is providing an "over-conservative" evaluated equivalent pressure.

The VIII Div2 model has in the background researches, FEA models and tests.
The equivalent pressure calculated by following this model must be more realistic.
It is a big argument to follow this model: it is now a model within a BPV AMSE code, not an external work.

2. When evaluating the EQP for piping flanges by following the VIII Div2 model, we can provide some margin for the unknowns considering, for example, 0.75m or less instead of m.

3. To limit p+pe as p_rating (the Kellogg approach) seems to be reasonable for piping flanges, since the gasket reaction is pressure proportional.
We might consider this limit for the total pressure applied on the piping flanges.

4. The ASME VIII Div2 relations "as are" (for the stress evaluation and J factor evaluation) and/or the Coade model can be applied for the corroded conditions, somehow in line with the piping codes intentions.

I understand the ASME Piping codes don’t intend to specify a method, for accounting the external forces and moments applied on flanges, and B31.1 Appendix II is just one acceptable method.
Mr. Breen and Mr Luf, your personal opinion on this matter (it would be negative…) shall be highly appreciated.
If you have a little time, please, what is your personal engineering opinion on a method based on ASME VIII Div2?
Thank you.

My best regards.

In this weekend I had some time to spend reviewing what I did.

In my previous post I said:

If calculated correct, the p+p1 total pressure shall produce exactly the same stress and exactly the same J factor and exactly the same flange rotation angle as ASME VIII Div2 predicts for p and Me and Fa loads.

Well, I have to recognize it is not the case with my previous result. The mistake is the load on gasket within 2007 ASME VIII Div.2 is specifically written for the design case. In fact it’s rather something elementary. Hmm…. That is, mea culpa. I’ve made the correction now (at least I hope so…).

I’m not interested to associate my name with a more accurate method but to use a correct relation for the EQP. That’s why this second draft is the last one. What I’ve put in writing is just an invitation for everybody to think about.
Obviously, there are a lot of intelligent people associated with this forum and able to do a better work. Please do it!
That’s why if you find out something valuable in my paper, you can use it with no restrictions.
Apart from my stupidity, some of my previous conclusions are still valid. The Kellogg formula is over-conservative; the pressure equivalence must be seen just a math substitution. The first sentence is well known, the second one I think is the only real contribution I had in this matter.


Probably, the only reason for which it is worth to use the pressure equivalent concept is the possibility to limit the total pressure to the rating pressure.
It would be interesting to see your opinion, and in fact the original question of this post was exactly this. Thanks, wromero!

About Kellogg formula. I said I have a deep respect for the Kellogg work and this is true.

By one hand, the classical Kellogg method was missed the factor FM. This factor just says “the flange is not a simple circular plate”. At the time of the Kellogg work, this factor was unknown. Today it appears to be a big progress. Thank you, Dr. Koves!

By the other hand, we are abusing on Kellogg formula by considering also the compressive force as generating a positive equivalent pressure. In reality, only a tensile axial force is reducing the load on gasket and consequently increasing the danger of leakage. A compressive force is helping the flange joint. Also, a torsion moment is not acting as a bending moment for the gasket stress, so such moments shouldn’t be mixed in the formula.

In the paper I’ve used some consistent ideas belong to Chuck Becht. Thank you and I’ve just hope I’ve understood well…

My best regards and my apologies for the mistake.

And finally I’ve just “succeeded” to post a wrong file. I try to re-attach the “right” version.
It’s clear I need your help, I need to finish my project and I need a long vacation!

I think the Coade flange calculator is more accurate than any theory based on assumptions.
In the same time, it makes sense to understand what really the equivalent pressure is, why it’s conservative and why it’s the preferred approach.

I’d like to share with you my conclusions.

1. The Kellogg EQP theory is a gasket load equivalence based on some assumptions.
The result depends on these assumptions.

2. We haven’t a clear understanding on the result. For example, a compressive force is "helping" the flange; anyway it’s not acting as a supplementary pressure that loads the flange. This is not clear in the formula interpretation.

In my first posted paper I’ve made a mistake assuming "mathematically" the gasket load as proportional with pressure. Now I guess this mistake has helped me to see what is wrong in our understanding of Kellogg formula. In fact, if we are looking only mathematically to the formula 2*b*Pi*G*(m*p) we can wrong understand that the stress on gasket is pressure proportional. Following this "understanding" we may say "under a bending moment and compressive axial force, the stress on gasket is increasing, so this is as an equivalent pressure is loading the flange". This is wrong because a pressure load is decreasing the gasket load and stress.

In my understanding, the Kellogg theory analyzes (in a specific way) the gasket tightness reserve, and for this reason the "tensile loaded part of the flange" must be considered.
Following this interpretation, the formula is the same, but now F is a tensile force and is acting –correct- as a supplementary pressure that loads the flange. Also the rating pressure as a limit for the gasket tightness makes more sense.
I’m attaching a paper showing "step by step" my interpretation.

3. This interpretation has helped me to understand why the Kellogg formula is so conservative: it’s a gasket load method presuming that the gasket is carrying on all the external loads.
Supposing we have only a bending moment load, this assumption is exaggerated for the tensile loaded part of the flange (and accurate for the compressed loaded part of the flange).
In fact, this is the worst scenario for the gasket load and the result is a very good formula for the Nuclear Code, where the worst scenario must be considered for the safety reasons!

4. In our days knowledge, I think that the Kellogg formula may be corrected by considering

M* [I/(0.3846*Ip+I)]*[G/(C-2*hD)] instead of M.

We would consider the correction... but is not included in a serious paper (or maybe is?).
You would get directly this result comparing the equivalent force in the Kellogg theory with the equivalent force considered by 2007 ASME VIII Div.2.

5. I’ve realized there is a "simple" effective way to create a realistic "by-pass" of the pressure equivalent method, in the same time estimating accurately the gasket tightness reserve.
But we can do this only numerically, case by case.

A structural model analyzing together the flange, gasket and bolts (as the Coade flange calculator is) can be used to find out a realistic minimum load on gasket.
"Minimum" refers to the fact the flange is under a bending moment and the gasket load is variable along the circumference, so has a minimum.

Having this result, we can estimate a simple "safety factor" versus the gasket load at rating pressure.
This safety factor may be defined as:

(minimum_along_circumference_ gasket load in operation)/ (2*b*Pi*G*m*p_rating).

This safety factor must be greater than 1; a safety factor of 1.2 means the minimum gasket load due to internal pressure and external loads is 120% of the value that occurs at p_rating.

In fact, this safety factor is:

(minimum_gasket_ stress in operation)/ (m*p_rating)

and is following exactly the approach of Rossheim, Markl and Taylor Forge’s "Modern Flange Design Bulletin 502".

For this reason, I think the best name of this safety factor is "the Taylor-Forge gasket stress index".
It’s a piping flange specific safety factor.

In my opinion, this index should replace the EQP method. The Kellogg EQP method makes (in a specific way) a similar estimation, but assuming the worst scenario for the minimum gasket load in operation.

6. Now few questions for Coade:
- is the HG force (calculated and shown in "Plot forces") the minimum_along_circumference_ gasket load
or
the maximum_along_circumference_ gasket load?
- is the axial force (introduced as load) a tensile force or a compressive one?
- it would be possible in the next version of this flange calculator to have either tensile force or compressive forces as loads?
-what is the meaning of the "Leak-Proof Joint (Creq)" in the Gasket Compression section results?
Thank you!

7. Last, about an EQP method based on the ASME VIII Div2. My first paper was wrong, my second paper (even correct in a way) is considering a particular gasket load and gives an ultra-conservative result. May be corrected and gives a realistic result, but I was the first saying it’s hard to be counted as a useful approach, just because it’s not more than an alternative calculation. The problem of the limits for this equivalent pressure is very questionable.
Instead, I’m very confident the Coade flange calculator makes the direct calculation better than any theory, because is able to estimate accurately the HG gasket load and not only this!

Best regards.

The Hg force is derived from the equation:

2 * b * Pi * G * m * P for the operating condition

Code:

Now:
            b = gasket effective width
            G = gasket effective diameter
            m = A pressure multiplying factor
            P = The pressure acting on the flange

The above equation is actually the equation for Hp, not Hg. See the post at the end of this thread for the correct equation for Hg. (R. Ay, Sept 29, 2008)

When an external force is being taken into consideration, the force is converted to an equivalent pressure (Pe)and added to the hydrostatic end force (H). It is not added to the force on the gasket (Hg). So, for the gasket force (Hg), that force is considered to be generated by the design pressure only, not the design pressure plus the axial force pressure. So, if I understand the question, the minimum or maximum load does have any bearing of the gasket force (Hg).



We use the absolute value of the force you enter, to make Peq as large as possible.



No, using a negative force would be non-conservative.



This is the required compression to prevent leakage.

Dear Mr. Richard Ay,

Thank you for your answer.

Really I appreciate it. For me is the first sign I'm spending my time with some profit.

Well, in this case I have some suggestions for the future flange calculator.

1. The Taylor-Forge/ ASME philosophy is focused on the conservative assumption that the pressure is discharging only the gasket.
The stress on gasket evolves from Y value at p=0 to the m*p_max value at p_max. The gasket load at p_max is HG=2 * b * Pi * G * m * P_max and is less than the gasket load at p=0 i.e. HG=W0.

For this reason HG=2 * b * Pi * G * m * P is a tricky equation when trying to describe the gasket load.
Instead HG=W0-P*Pi/4*G^2 is the correct equation ("correct" in the limits of the assumption made). The equation is more complicated when a bending moment is acting, because in this case HG is variable along the gasket circumference.

I understand the flange calculator is a structural model based on correlations on how really the bolts and the gasket are interacting.
For this reason I’m sure that you can calculate- numerically- the HG load given by the coincident p_operation, M and F, by following this model instead to consider an equation to describe it. In this assumption, the "Taylor Forge" index I’ve proposed makes sense, because in operation we need to be sure that

HG_operation > 2 * b * Pi * G * m * P_max.

This index would be:

Index= HG_operation/ (2 * b * Pi * G * m * P_max)=
=(Gasket stress in operation)/ (m * P_max)>1

No code is asking this index because no code is checking the operational conditions.

2. About the contribution of the axial force in Peq.

It’s true that considering negative values for forces (compression) it’s not conservative for the design of the flange. This is specifically asked in ASME VIII Div.2.

But a compression force might be a real condition for operational conditions applied on a piping flange.

The equivalent pressure concept is conservative enough because the assumption made, it’s no need to help this formula by considering axial force as tensile forces if we really have a compressive force.
For this reason, really I think in the next version, the calculator would accept negative (compression) forces.

Thank you very much for your information...and of course for your calm endurance considering my long posts, getting bored everybody.

My best regards.

Dear Mr. Richard Ay,

1. If the Coade flange calculation is based on a formula for the gasket load, I suggest you the following:

HG=W0- P*Pi/4*G^2- F- 4*M*[I/(0.3846*Ip+I)]/(C-2*h_D)

where F is the tensile axial force, M is the bending moment, the rest of the symbols as per 2007 ASME VIII Div2.

You can get this result after drawing a sketch with the forces involved in the moment equation of ASME VIII Div.2. The forces equilibrium equation gives the gasket load, presuming the bolts load remains at constant value W0. I attach a paper including such sketch. In the same paper you can see two ways for the equivalent pressure calculation, but I guess nobody is interested on this subject. Good news, it’s really the last paper!

The gasket load equation you’ve considered it’s a "tricky" one. It works for pressures near rating pressure, it doesn’t work for low pressures- being far conservative for the flange moment calculation. An effect is the flange calculator gives unrealistic results for the ASME Rigidity Factor "J", Seating Case (in the case I’ve checked was 1.7). Obviously, the piping flange hasn’t a problem in non-operating conditions. I think you need to correct this anomaly, and the above equation is giving realistic results.

2. I suggest Coade to improve the "Flange Leakage/Stress Calculation" description in User’s Guide, at least some clarification on page 12-20 will be high appreciated.
In this moment the description leads to the idea the calculator is based on the FEA methods correlations, that’s why I’ve presumed the gasket load is numerically calculated.

3. The bolts load constancy assumption (i.e. bolts are not participating, always W=W0) is very good for the design case, where the bolts have "no credit" to interact with the flange loads.
For the flange checking case, probably this assumption is exaggerated, and I think a structural model based on FEA researches must replace this assumption.
I hope Coade shall be interested to develop such flange calculator and it shall be possible to evaluate more accurate the reality.

4. Until that time we are obliged to follow the pressure equivalent calculation (that is a gasket load equivalence maintaining the conservative assumption above explained).
I second Mr. L-C Peng opinion "it is so conservative that it would probably disprove most of the installations which are operating satisfactorily"- see the article "Evaluation of Flanged Connections Due To Piping Load".

In fact, today I think we really need to check the gasket load vs. the pressure rating gasket load value; the only issue is the way we are counting – as EQP- is leading to a conservative result. A numerical approach probably shall relax this condition.

Well, this is the end of the road for me. It would be a beginning of a new approach for you and I hope my suggestions are helpfull. Might my relations be correct or not so correct, I think more important for all of us is to progress in our knowledge just to have more improved results in practice. A part of this progress should be to understand the classical results – as the Taylor-Forge approach and Kellogg’s EQP are- and to improve them using the today’s tools. I’m very confident this will be soon the case of the future Coade flange calculator.

Have any of you read the original paper? It states that this method is on ly for a quick check and if the system passes then you have a very conserative design if it fails then you need to look at the flanges with more accurate methods. I recommend you read Section III, Division 1 - NC-3658.3.

You are right, of course.
Yes, in some cases we need to look at that piping flange with more accurate methods.
And which method is practical for the piping flanges?
I expect the next flange calculator will be exactly this tool. That means a tool that not just guesses the gasket load but calculates numerically it. If this numerical result shall be in range of 10-15% of the theory result, that means also the theory is good. If the difference is 50% it’s a complete other discussion.

NC-3658.3 says: The pressure shall not exceed "X" times the rated pressure. What X do you consider for a piping flange which is not under NC-3658.3 requirements?

Best regards

The equation for HG is:

Dear Mr. Richard Ay

Thank you.
New information, some new questions for Coade. Thank you in advance.

1. If I understand well, do you intend now to proceed with a HG formula based on ASME VIII Div1? Or maybe is this formula already implemented in the Flange Calculator?

2. The ASME VIII Div1 formula doesn’t include the external loads. It’s just a design formula that doesn’t count such loads. For the flange checking purpose, your approach is to count the HG as affected or not affected by the external loads? If yes, can you please give some details?

3. I understand from the formula posted now, the calculator counts the design bolt load as the ASME VIII Div1 Appendix 2 requires. However, this is a design rule.

It seems that some B16.5 flanges deviate from this rule. That isn’t an "incompatibility" between the codes/standards, since the Appendix S of the ASME VIII Div1 says clearly:

"In any event, it is EVIDENT that an initial bolt stress higher than the design value may and, in some cases, MUST be developed in the tightening operation, and is the intent of this Division that such a practice is permissible, provided it includes necessary and appropriate provision to insure against excessive flange distortion and gross crushing of the gasket" (I've just put two words in capital letters).

That explains why, in some cases, the result of checking shows "weak" bolts while, in reality, they are developing adequate bolts load.

Anyway, is the Flange calculator counting this reality? That means probably another rule for W, correcting in some cases the design formula by forcing a more realistic result, exactly as the Nonmandatory Appendix S suggest?

Thank you again,
My best regards

Mariog,

You quote from Appendix S, and quite right, residual bolting stress will usually be far higher than the design stress value, and in reality it must be.

The incompatability I referred to in another post does not concern weak bolts. The common case of bolt-up stress exceeding the allowable when operating stresses re acceptable generally results from a high radial stress.

Using a higher than 'design stress' value for bolt allowable stress will only make this condition worse due to W = 0.5(Am+Ab) x Sb. However it's not a real problem as we all are aware.