Fatigue Analysis of Piping System

[gimini21]

To all experts out there,

I have modeled and run a jacketed tubing system running on 20 - 110 deg F. The temperature difference of 90 deg. F can result a thermal cyclic change everyday. I am asked to perform a fatigue analysis after I run the model whith fairly good stresses anywhere, no overstressing. My first question is, does an element going to experience fatigue failure due to cyclic temperature change even if such element is within allowable stress from B31.3? Secondly, I am using Caesar 4.2 and seems like it does not have the capacity to peform fatigue analysis, and so is there any method of peforming this type of analysis say by manual calculations? I would appeciate it if you guys out there could give tips and resources on performing this calcs. Thanks and more power to this forum.

Gimini21

 

[prex]

There are at least two more factors you need to consider:
1)Fatigue failure is caused by peak stresses, and these in turn are determined by the details of your design, e.g. how the jacket is connected to the pipe (fillet welds, full penetration welds, forged Y pieces?). The evaluation of peak stresses involves the use of the stress concentration factors or of detailed FEM models. Some details (e.g. fillet welds) should be even totally avoided in a high fatigue environment. It would be surprising that peak stresses stay below the allowable stress.
2)You need to define a load cycle to evaluate fatigue, and the governing quantity will be the stress change (or range) during the cycle. If the thermal expansion is the only changing load, then your cycle is from cold to hot temperature and the stress range is the stress intensity variation from cold to hot.
A method for evaluating fatigue is given ib App.5 of ASME VIII Div.2. The allowable alternating stress (this is half the stress range as above) at 10000 cycles is of the order of 20 ksi, so you shouldn't be in a bad condition (but of course you must include peak stresses in your calculation).

prex

http://www.xcalcs.com

Online tools for structural design

 

[LSThill]

gimini21 (Mechanical)

Post your question to CAESAR II FORUM:

http://WWW.COADE.com

Leonard@thill.biz

http://www.thill.biz

 

[gimini21]

My sincerest thanks to you Prex. Though this is a tough job for me I wanted to make a try and have this advice on my side (already on my thick clippings right now). More power to you and thanks for the advice too Isthill.

Gimini21

 

[LSThill]

Hi gimini21 (Mechanical)

I use CAESAR II to set up the modle and check the Fatigue Analysis of Piping System using FE-PIPE.

A.W. Paulin Jr., P.E.the original author of the CAESAR II pipe stress program, and coauthor of the FE/Pipe finite element program and the BOS Fluids, transient fluid simulator.

Leonard Stephen Thill
Leonard@thill.biz

http://www.thill.biz

 

[gimini21]

Isthill,

Please let me know something about FE-PIPE. Is this software affordable to own personally? I mean Caesar is expensive on license and then you owe $12 per analysis. Thanks for the good info here. Regards.

Gimini21

 

[LSThill]

HI gimini21 (Mechanical)

Paulin Research Group News Issue 03.01Software Releases: The new versions of all Paulin Research Group and Dynaflow software packages are nearing completion. Quality
Assurance documents have been prepared for BOS Fluids and FE/Pipe. Final checking of AxiPRO and new NozzlePRO
features is underway. Watch this space for availability.
Seminar: The last three day FE/Pipe, NozzlePRO and AxiPRO seminar of this year will be conducted November 10-12 in Houston.

New features in all products will be discussed. New technologies for predicting flange leakage and for calculating stresses in nonintegral repads for saddles and pipe shoes will be covered. The course will include the updated 2003 format. Basic-to-comprehensive FEA technologies for the PVP industry will be covered. The course will be taught primarily by Mr. Tony Paulin, the original author of CAESAR II, and the coauthor of FE/Pipe, NozzlePRO and AxiPRO.

Contact sales@paulin.comfor more information. Laboratory News: The installation of the Markl tester is about complete and flange leakage testing with helium is underway. Liquid leak rate correlations are published in the Paulin Research Downloads section, and have been installed in the upcoming version of AxiPRO. Electrical wire heating, surface tensiometers and micro-hardness testers have recently been installed.

The photographs below show Kent Stephenson of High Tech Fabrication installing the Markl test frame, and James R. K. Sauseda of MK Supply conducting a helium leak test.
Markl Test Rig Installation Helium Test Bolt Tightening New Features: There are numerous new features in each package, but a few of the more interesting features include:

1) Thermal Head (FE/Pipe) for bowing in horizontal and sloped pipelines and transfer lines.
2) Saddles and Pipe shoes in NozzlePRO/FE/Pipe. (Integral and Non-Integral repad models.)
3) API 618 Compliance for Reciprocating Compressors in BOS Fluids.
4) API 674 Compliance for Reciprocating Pumps in BOS Fluids.
5) Slug and Flashing features in BOS Fluids.
6) Gas Transient Analysis for Relief Expansion of Non-Ideal gases. (BOS Fluids)
7) Automatic 3D Modeling of Nuts, Bolts and Flanges with Leak Predictions correlated with laboratory tests.
(AxiPRO/FE/Pipe) Analysis options include torsion, external moments, and bolt interaction analysis. ASME App 2, BFJ
and EN 13445 Annex G calculations are performed with each FEA analysis.
8) New Security System ­ Permits transfer of license and web activation.
9) Earthquake and Pitch and Roll Sloshing Base Shear Loads and Overturning Moments (SPLASH) 10) Large Diameter Pipe Modeling (FE/Pipe)
11) Brick Models of 4-way Tees and WFI/Bonney Olets Saddle Option in NozzlePRO Spillover Analysis in Splash API 618 Compressor and Structure Vibration Analysis Left-to-Right: James R.K. Sauseda, Patrick Marcotte, Mary Jane Metcalf, Tony Paulin, Kristin White, Hans Bos

 

[gimini21]

Isthill,

Thnaks for all these informations. Great news for everybody I believe who hope to learn more on this fiel like me. More power to you.

Gimini21

 

[Jsquare]

If you're using B31.3 piping code see the section about Stress Range Reduction Factor. If you're only cycling once a day and designing a for 20 yr plant life that's only 7300 cycles. Assuming that stresses are low, I feel it's not worth evaluating for fatigue since there will be some shut-down time as well. Plus, you probably need to provide more info. as to what your problem is and where. Are you concerned about fatigue of sypders or the wall?

 

[davefitz]

prex is basically right. A 90 F daily change in temperature is no significant for normal metals fabricated with an eye on reducing stress raisers, but if you have no prior experience with the design in the intended service, then you should conduct a FEM of the likely stressed areas.

Some minor improvements such as full penetration welds in lieu of partial pen welds can greatly extend fatigue life.

Also, for a 20 yr life, corrosion fatigue interaction may be the governing failure mechanism , in particular if the welds contain pre-existing cracks that harbor corrosive elements ( partial pen welds, for example).

 

[bvi]

As it is a jacketed system, you will have components that are not covered by the sif factors in Appendix D of ASME B31.3. So, finite element analysis (or other calculations) may be appropriate to determine what stress intensification factors apply.

Another consideration is that in jacketed piping systems, the piping is self constrained. That is, the inner and outer pipe restrain each other via axial forces. These axial forces are not included in the Code stress calculation. Look at the operating case in CAESAR II to get an idea as to the magnitude of these stresses. CAESAR II includes them in the operating stress calc. There is a sentence in B31.3 that warns about this effect. If these axial forces are signficant (and they will be in a jacketed system with the pipe and jacket at different temperatures) they need to be included in the evaluation.

Another effect is elastic followup. The thermal expansion stresses on the component joining the inner and outer pipe must be designed to be less than yield stress, or else it will be subject to strain concentration. Once it starts to yield, the pipe will act as a spring and keep pushing it. In this circumstance, the strain range in that component will be higher than you have calculated on an elastic basis.

 

[gimini21]

To Davefitz and Cb4:

My sincerest thanks and appreciation to your suggestions though this thread has been posted for a while. The idea of going into a FEA on the specified elbow that the client is very concerned about was also suggested by somebody to me working on FEA (aerospace field). The fluid here is highly hazardous and also the project used tubing which is basically thin though SS and double contained. Yet corrosion as Davefitz mentioned became another key design criteria that was strongly suggested to me too. Cb4 ppoint of axial stress for the inner pipe is another interesting thing I learned. I have actually sent the model to COADE and have it checked specially how the inner elbow is restrained to the outer elbow. And so to proceed into FEA, streses from CAESAR must be well within the real world. My model is okey and right now, I am close to using ANSYS 6.1/Ed to do a specif FEA on a particular elbow. (I've been studying this FEA using ANSYS for the last few months at home - not so good in DE on my school days so it take a bit time to grasp even the basic). You both are smart and well-experienced. Regards and more power to you. Thanks for the time.

Gimini21

 

[LSThill]

Hi gimini21 (Mechanical)and Team members.

Note: Regards to ANSYS and FEA Analysis in Houston 1998 the Bench Mark between ANSYS and FE-PIPE, in the Process Petro-Chemical, Oil and Gas: On-Shore, and Offshore regards to Piping and Pressure Vessels FE-PIPE, Yes!

I have been using ANASYS sence 1988 but I love: (FE-PIPE / Nozzle PRo / All PRo).

Leonard@thill.biz

http://www.thill.biz

 

[richay]

Fatigue capabilities (including "cummulative damage&quot were added to CAESAR II for Version 4.10.

 

[bvi]

Fatigue analysis and cumulative damage are included in the B31 Code rules. The f factor that reduces the allowable stress at higher numbers of cycles is based on fatigue testing of piping. Cumulative damage is taken care of by calculating an equivalent number of cycles. In the end, the maximum stress range is compared to the allowable, and the number of cycles considers all conditions. Conditions at lower stress ranges than the maximum are converted into equivalent numbers of cycles at the highest stress range. It's all in the code rules.