Elastic follow-up in high temperature piping.

[antoinecom]

I am interested in obtaining a design method for analysing the problem of "elastic follow-up" in high temperature piping. An elastic analysis shows the thermal displacement stresses and sustained stresses to be well within the B31.3 allowable. However, cracking has occured in certain regions and metallugical examination does not indicate evidence of the usual suspects.
Is there a simplified approach that does not involve non-linear fe analysis.

 

[DSB123]

antoinecom,
Where are the cracks located? Could it be due to poor PWHT? (or lack of it) Also maybe the SCF's are the cause!!. There are a number of reasons why cracks could be occuring. What's the material? Temperature?

 

[antoinecom]

The material is ASTM A335 P11 operating at 15 barg and 500 Deg.C. All the PWHT specified was carried out. The cracking is in an area of stress concentration where a manifold is connected to furnace tubes.The central question is concerned with "damage" caused by thermal cycling of the piping system when a major part remains "elastic". It follows that "strain accumulation" in plastic regions does the damage.

 

[JohnBreen]

Hello,

1 1/4 Cr 1/2 Mo material has become famous for creep-fatigue interaction. That is why the allowable stress was reduced some years ago. Cracking usually occurs at branch connections. 900+ degrees F is a lot of temperature for this stuff - probably there will be some carbon migration to the grain boundaries and resulting embrittlement. Use A-335-P22 next time.

I think Dr. Charles Becht covers the elastic follo-up issue in his book but I do not have it here in Virginia to check.

Regards, John.

 

[rmw]

Is this actually external piping, or a header associated with a furnace or boiler??

rmw

 

[davefitz]

If it is a header in the boiler, there will be a tube to tube temperature unbalance such that some tubes ( so called "worst tube") will operate at a temperature higher than average, and the ligament region of the header near the worst tube will then operate at a temperature higher than the header design metal temperature and experience accelerated creep damage.

Also, the P11 header is likely a lower alloy than the tubes, so there may be a dissimilar metal weld at the tube to header weld ( bad design if true).

Some headers have a geometry such that all tube to header welds occur on one side of the header; this implies the header will bow or hump during startups and cause excessive loadings at some parts of the header. If the loading is greater than yield ( locally) there will be generated residual stresses which are then creep-relieved at steady elevated temperatures .

Several design methods to avoid these problems are :
a) stub to header weld to be of same alloy
b) stub to hdr welds to be aligned on 2 opposite rows for symetrical heating of header, or else support header such that bowing or warping will not result in increased stresses. Examples may be to have the stubs weld into the side centerline since horiziontal warping can be easily accomodated by vertical support straps, or else use constant load spring supports.
c) use a higher grade alloy which has much higher allwoabel stresses at elevated temperaures such that secondary stresss do not significantly contribute to creep damage.

 

[antoinecom]

Gentlemen, This furnace manifold was external to the furnace and there was a dissimilar weld connecting the tubes. However, the welds were sited away from the manifold connection and did not exhibit cracking. The reason I thought elastic follow-up was the reason for the cracking was that elastic calculations showed very nominal stresses and the cracks had appeared after four years operation. This indicates less that 100 cycles of operation.
Thankyou gentlemen for the information.

 

[bvi]

You could do a creep analysis of the piping system but I presume you may not have the tools for that. With a typical piping flexibilty analysis program, you can do an approximate analysis to get a feeling for the problem. The method is described in the paper "Elastic Follow-up Evaluation of a Piping System with a Hot Wall Slide Valve," C. Becht IV, 1988, in PVP-V0l 139, Design and Analysis of Piping, Pressure Vessels, and Components - 1988, ASME.

Run the analysis, and then rerun it with the local area you think is subject to elastic followup with a reduced value of elastic modulus. This roughly simulate accumulation of creep strain. Then look at the change in stress relative to the change in strain. If there is no change in stress, it is 100% elastic followup. If the change in stress reduced in accordance with the simulated creep strain increase, then there is no elastic followup. The simulated creep strain is the the stress divided by the difference in the elastic moduli used in the first and second analysis. The analysis can be repeated with various values of elastic modulus.

 

[bvi]

Didn't state that precisely right. The change in strain is the stress from the 2nd run divided by the elastic modulus for run two minus the stress from the first run divided by the elastic modulus for the first run.