B31.3 P&T Event Duration for Flexibility Design 1

[KANN]

For flexibility design I interpret B31.3-2004 to require the design pressure & temperature to be the most severe condition of coincident pressure and temperture that can occur, and is not tied to the duration of the event. (The event would obviously have to have a long enough duration to heat the pipe to the expansion temperature under consideration.) Flexibility design is, however, tied to range and frequency of the event, with respect to displacement stresses and fatigue.

Pressure design (Para. 302.2) is tied to duration of a pressure and/or temperature event, and specific provisions are in the code (Para. 302.2.4) for occasional variations in temperature and pressure to meet pressure design requirements.

An alternate viewpoint, by the owner, is that what is important is to design the system to the temperature that occurs, say 99.7% of the time. A higher temperature for a "short" duration does not matter.

Since steam delivery contract limits state a higher upper limit to steam temperature, it really does not matter to me that the plant has not recorded temperatures at that contract limit for the last several years. (Is the monitored temperature being correctly measured?) What I propose is using the limit of the steam delivery contract as the design temperature, which is viewed as overly conservative by the owner.

Comments on the code interpretation or the view of conservatism?

 

[StressGuy]

In most applications that I've dealt with, the "design" conditions are intended to apply to the wall thickness calculation and usually have padding in them that makes them inappropriate for flexibility analysis.

I would suspect that your "contract limit" is going to be a similar number.

In reality, code compliance is the owner's responsibility and they should be providing you with the conditions to consider, since they should know best the conditions of streams within the facility.

In your case, I would probably run both temperatures and see how much, if any, additional flexibility would be required for the higher temperature or if you could justify it by using something like the "liberal" stress range allowable.

Edward L. Klein
Pipe Stress Engineer
Houston, Texas

"All the world is a Spring"

All opinions expressed here are my own and not my company's.

 

[KANN]

Edward -

Thanks for your thoughts.

Since my initial posting I decided to use the upper temperature limit of the steam delivery contract for the flexibility design. This decision was based on my interpretation that the duration of the temperature event is not a consideration for flexibility design. If the temperature could reach the contract limit, even for a "short" duration, the piping flexibility needs to be based on that temperature. Yes, the contract limit probably has some padding added, but I'm not going to try to tease out what the steam supplier thinks is real and what is padding...I'm going to go with the stated limits of the contract.

Recently, there was a local power outage that resulted in a pressure surge and some loss of control of the desuperheater station at the generation plant. This upset resulted in SRVs firing at a number of our client's pressure reducing stations. Some of the piping at reducing stations moved more than anticipated resulting in steam flow exiting SRV vent pipes at drip pan elbows. This seems to have gotten the owner's attention (what will probably never happen or has hardly ever happend...happened) and using the higher temperature limits for flexibility design has not been an issue.

 

[davefitz]

A complicated issue, for sure.

There are different failure modes, so the answer may need to be specific for each failure mode.

For damaged that is worsened by creep , then time durations on the order of hours might be needed ( for creep relaxation and associated ratcheting and pipeline growth, when the primary stress is related to pipe growth )

For other time dependent issues, such as creep with fluid pressure as the primary stress or for graphitzation, a time period of weeks may be required before a real techical concern developes, although the codes maight not recognize this officially. For example, B31.1 will not allow a pipe to be overheated above the max temperature for which an allowable stress is published in the code, even for a second. This is a problem for carbon steel overheated to 801 F.

For purposes of determining if the pipe itself has reached the fluid's overtemperature, I believe the code recognizes that the temperaure of record is the pipe's mid wall metal temperature, so very short transients shorter than t^2/a could be neglected ( t= thickness, a= thermal diffusivity)since the midwall has not yet reached equilibrium with the fluid. For pipes handling liquids that have a low convective heat transfer coeficients ( ie oil), the biot number (ht/k)^-1 can be used as indicative of the important time interval.

Some pipeline codes are based on the assumption that the max stresses will not exceed some multiple of allowable stress for not more than ( 1000 ?) cycles. So, it seems you would need to ascertain first the number of cycles the event will occur, and if it is less tehn (1000 ) cycles, then adequate flexibility is added to keep the max stress below the code defined allowable.

Significant overheats will occur at many large power plants and at complex chemical plants. Certainly, it is conservative to build enosugh freedom of growth ( to fixed bumpers) to ensure the solid bumpers are not contacted until the overheat is at least 50F over the max you could normally expect. Once teh bumpers are hit, the pipline stresses increase dramatically.

 

[JohnBreen]

Interesting,

You might benefit by going to Chuck Becht's site and carefully reading the article on the latest B31.3 changes. Read it all.

http://www.bechtengineering.com/processPower/articles/ASME-B31.3TechnicalChanges2004.htm

And there are many useful articles at this site so look around a bit.

Regards, John.

 

[KANN]

Davefitz -

I'm not dealing in the creep range in this instance...yet. Found your discussion on time to temperature helpful in understanding approaches to take in calculating mid-wall temps.

John -

Thank you for the reminder...I hadn't been to Becht's site since January and found it helpful - again. Becht & Diehl's ASME paper on the 2004 code changes is a good read too.

- Ken