Piping Stress analysis & Flanged Joints 2

[edemarcos]

Often, when I perform a stress analysis of pipe system ( B31.1 or B31.3 ) I have that at the flanged joints I have a lower than allowable stress.
If I calculate the equivalent pressure Pe at the flanged joint location I have the ( Pd+Pe )> Prating.
Due to the fact the ASME B31.1 and 3 do not require the pressure chek at the flanged joint, I suppose the line, and the joint, is/are OK.
Or not ?
many thanks for Your help

 

[bvi]

You cannot assume you are OK because the stresses are within the allowable stress for the pipe. As the designer, you are also responsible for making sure that the loads on the flanges will not cause them to leak. See for example 319.1.1 "Piping systems shall have sufficient flexibility to prevent thermal expansion or contraction.......from causing.....(b) leakage at joints."

However, no specific rules are provided within B31.3 for assessing this. The equivalent pressure approach is highly conservative, and will fail many flanged joints that are actually acceptable.

One of the reasons (and here 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.

 

[LSThill]

edemarcos (Structural)

I agree with cb4 (Mechanical)and additional Paulin Research Group News Issue 03.02Flange Leak Testing Continues
Cold creep of the composition gasket currently being tested is increasing the leak tightness of the joint
each cycle, well past anticipated values. Testing is in its 13thday and the increase in number of seconds for a 100 psi drop in pressure is shown in the plot below.

The gasket is cycled from the seating stress to the test stress and then helium introduced at 500 psi. The
plot shows the time it takes for the pressure in the rig to drop from 500 psi to 400 psi.

AXIPro

AXIPro is a new product offered by PRG. AXIPro is a powerful, easy to use axisymmetric and brick finite element modeler. Bolts, nuts and holes may be included in the 3d brick models so that users can see the results of hole spacing on the stress distributions. Dimensionally accurate flange models including studs and nuts are generated automatically for six major flange standards. Analysis results include graphical representations of ASME stress intensities, resultant bolt axial loads, gasket stress distribution, overall displacements and flange separation. It allows automatic 3D models; App2, ASME BFJ, and prEN 13445 Code evaluations; overturning moments, shear loads and torsion; gap elements and initial tightening; interaction analysis (how does tightening or loosening one bolt affect others, and more importantly; how does it effect gasket stress); and leakage prediction PPM, pound/year, and drops/ hour. User’s can include dished heads with body flanges, ANSI flanges, etc. AXIPro has its own user’s guide and is included as a module in FE/Pipe.


edemarcos (Structural) if you required any additional detail information please contact

Tony Paulin,

http://www.paulin.com

11211 Richmond Avenue Suite 109 Houston TX 77082 USA Ph: 281-920-9775 Fax: 281-920-9739


Leonard@thill.biz

http://www.thill.biz

 

[tonycola]

I agree that if you use the method that calculates the equivalent pressure and then compares the result to the flange rating, this analysis is very or perhaps too conservative.
A less conservative, but acceptable approach would be to calculate the equivalent pressure and then used this pressure in the formulas of ASME section VIII, div. 1, appendix 2, to calculate the flange stresses. These stresses should then be compared to the allowable stresses for the flange material.

Good luck.

 

[Elvie]

I also agree with cb4;if for piping under ASME B31.1 a leak may be disastruous under ASME B31.3 a HC detecting system (in Oil & Gas Plants) will help in quickly healing the leakage.What is sure is when you calculate a bolted flange connection as per ASME VIII Div 1 App 2 in most of the cases is the Wm1 ("bolt load for operating conditions&quot the governing factor in connection sizing.Wm1 is the sum of H ,needed to face the pressure effect (hydrostatic end force)and Hp ,needed to keep the gasket set .Both H and Hp are based on P= max allowable prssure, which in most of the cases is the pressure rating of the relevant piping class (i.e. of the ASME B16.5 flange rating).So the bottom line is that flange connection is already sized for loads well above pressure rating of the flange.
So as cb4 says pretension the bolts at +/-50% of their material SMYS if you use flange connecton (although I prefer the use of hub connectors so the problem is not more there).
If some one goes for deeper investigation in a high stressed flange connection I will guess he's going to find the weakest point at the flange WN butt weld to the piping on both sides of connection.

 

[LSThill]

edemarcos (Structural), cb4 (Mechanical) and TEam Members


During a recent analysis of body flanges that sandwiched a tubesheet for heat exchangers, engineers had attempted to use software that did not seem to be capable of handling flange pairs with differing conditions as described in Appx. 2 - 2-5(c)(2) of ASME Code.

A typical procedure to include these requirements is shown below:

1. Calculate Wm1 & Wm2 for each flange (side of the tubesheet)
2. Determine which load of the four governs the minimum bolt area. The side of the tubesheet that governs the required bolt area becomes the “independent” flange and the other side is the “dependant”.
3. Set the actual bolt area and calculate W.
4. The subsequent calculations for the independent flange are carried out as normal.
5. The subsequent calculations for the dependant flange are also carried out as normal except that the value for “W” is the governing value from the independent side. (This affects Hg).

Note: End users will at times require extra bolting or the W to be calculated at actual bolt area rather than average or many other variations.

It would be reassuring if software of this type printed warnings to it’s limitations so that they would not be accidentally utilized in an incorrect application.

A). With the introduction of the ASME UHX rules the flange\tubesheet interaction gets even more complicated than it was under TEMA. Note that Coade served on the UHX committee and I'm told that their software was used to produce the "validation" examples seen in UHX. One consequence of this can be seen in the tubesheet examples in UHX where problems in the calculation appear. For example, the flange load W actually varies WITH EACH LOAD CASE (note there are 7 load cases to be investigated per design condition) and yet the flange load W is taken as a constant across the 7 required load cases! Also, the flange load W used in each example is un-realistic in that it's not possible for ANY real flange to produce the stated bolt loads using the example pressures. I suspect this is because the bolt load is a single input into the Coade HX module and so the tubesheet design is not completely linked to mating flanges in their software. Worse still, any changes that affect the flange require the user to remember to come back and revise these tubesheet inputs.

BTW we've advised the UHX committee of the issues in the examples.

lst

 

[ScottMyu]

Mr. L. Thill

You are certainly correct that the new rules in ASME Section VIII Division 1 Part UHX are generally more complex than those in the TEMA standard. You are also correct in your understanding that COADE does have ASME Code Committee involvement in both the ANSI B31.3 arena as well as the ASME Boiler and Pressure Vessel Code areas. It is also true that as a company we dedicate our time and resources to further the development of these Codes and Standards. However, I feel that you are most incorrect about your statements regarding the example problems that appear in Part UHX. Firstly, the example problems were not generated by any of our software products. The examples were generated using MathCAD and Excel spreadsheets (W is not a computed value in these spreadsheets, but a direct input). Furthermore, COADE had no involvement in the generation of the software used to create the UHX samples in the current Edition and Addenda of the Code. The Special Working Group on Heat Transfer Equipment (SWGHTE) is composed of several members who have extensive knowledge of pressure retaining equipment and heat exchangers. Your comments regarding Part UHX and COADE are not only subversive and ill advised, but are also offensive to the untold collective of hundreds of individuals that have sacrificed their personal time for the betterment of the ASME Code. You may not realize that it has taken a tremendous effort on the part of many members to get the Code rules where they are today.

Let us now examine the statement that you made above regarding the flange design bolt load W. You seem to imply that for an ASME tubesheet calculation the value of W should vary from load case to load case. It is true that when performing a typical flange calculation W is to be calculated for both the gasket seating and operating cases and the most severe value used. However, you need to actually read the ASME Code Part UHX to determine the required value of W for use in the equations that require it. The definition of W as it appears on page 301.16 (2003 addenda) is as follows (similar in other places in UHX as well):

“W = channel flange design bolt load for the gasket seating condition, lb (N). Use formula 4 of 2-5(e) and see UHX–4(c).”

So, it seems as if the Code requires only the gasket seating condition to compute W. Now, if we can show that W for the gasket seating case does not involve the pressure term P, then it stands to reason that W will remain constant for all loading cases, even those that do not involve the pressure. From Appendix 2, 2-5 (2) (gasket seating): “The conditions existing when the gasket or joint-contact surface is seated by applying an initial load with the bolts when assembling the joint, at atmospheric temperature and pressure”. Sounds to me like this condition does not involve the design pressure (and I think we all already knew that). From that we can conclude that W should not vary between the load cases. In conclusion, it seems that your above line of reasoning is flawed and you need to reconsider your statements.

Regarding our Vessel design software, it is up to the user to insure the correct values are typed in before performing the analysis. There are merge features that allow the user to easily update values from other components including flanges. It is interesting that you are making comments and claims regarding our vessel software while at the same time you do not seem to be a registered user (Thill Eng). Additionally, in the future if you have questions about our company, ASME involvement or software, kindly call us or stop by our offices and we will be glad to help you. If my memory serves me correctly you have been to the COADE offices in attendance of our CAESAR II Pipe Stress Analysis seminar.

Best regards,

Scott Mayeux