Question for all the B31.3 guru's out there.
I have a packaged unit I am working on that has PSV's fitted specifically for relieving pressure during a fire [set at 47barg at 267C]. Normal operating conditions are 30barg and 80C.
B31.3 says very little about design to survive a fire, but on the basis that it is at least as likely to occur as an earthquake or extreme wind condition I am looking at checking the system over at 47barg & 267C and comparing the results against the B31.3 OCC stress limit @267C, as ideally we want the system to remain intact [if a bit charred] after a fire and not propagate the incident by further failures.
Since there is also significant thermal growth expected here - but probably only for one cycle, do I need to check the displacement stress range? I have a feeling that the system will be overstressed when comared to the allowable displacment stresses, but want to do the 'right thing'.
Also if we assume that we can increase the nozzle allowables on equipment by 1.33 we will proibably overload them as well during a fire.
Am I being overly conservative here? To me if we have a specified fire over-pressure protection, other safeguarding precations must also be applied.


Any comments, guidance or thoughts anybody?

The parameters of what you are trying to achieve with the fire case is the main problem, as one engineer says in this firm, 'tis a bit of a fudge'

for us, the fire case is about ensuring flamable vapours do not escape from the pipeline during a fire.

(We) assume that the system will be destroyed during the fire and that that code stresses are largely irrelevent.

The pressures and temperatures you are designing for will mean a difficult and problematic analysis, not to mention very strong pipework.

we try to engineer the problem so that the fire requirements are clearly defined and that excessive pipework design is kept to a minimum.

this can sometimes mean the de-rating of the PRV's.

Normally though, this whole fire case buisness is limited by the gaskets within the system, and is analysised accordingly.

Hint, Fire is not an operational case.....

So I presume that the hint has provided adequate help on this matter???

B31.3 really does not cover this case as far as flexibilty analysis is concerned. The closest thing that comes to mind are limits in Section III for SSE I suppose.

One could use f=1.2 I suppose for the calculation of SA for the thermal displacements, go from there. But if your going to do this sort of thing I would make sure the pipe supporting elements would be rated for the fire as well.

John
Thanks - I take your point about you beleiving fire not being an opertional case, but in the back of my mind I still have two concerns.
During an earthquake, blast, or other more 'usual' ocasional operating scenario, we wish the piping to stay intact, not to bust supports, rupture or do anything that might lead to further loss. Therefore as I understand it the intention is to maintain the structural and pressure integrity of the system and not nessecarily provide 100% operability after the event. Likewise I believe the same should apply to the 'fire case'.
If I have marked on a P&ID, that a PSV is designed for pressure relief during a fire - I must assume that this is a likely scenario, and we would wish the piping to still be there to give the PSV somewhere to vent safely to.
B31.3 para 302.3.6 is about 'operation' and does not say anything about normal operation or what does and does not fall within this scope.
This is now a relatively moot point personally, as my system is proving to be acceptable at the stated PSV stated relief temperature [as simplified as this is admittedly].
I do hear what both you and Superpiper are saying and I am not going to get too worked up about this and I am glad there is a forum where concerns can be raised, ideas can be bounced around and opinions given.

Hi Captain Kenny,

I have more or less the same problem as you in as much as my design temp is 235C at 7 barg and flex (fire) temp is 350C at 7 barg.
As I see it the code already covers this in para 301.3.Since the fire case has the highest temperature , I have treated it as my design temperature and made the system code compliant.
I think you are right about treating the growth as a 'single' cycle and keeping the expansion stresses within the stress range.
The fire case is not like the blast case; in the blast case you would have to design to say 1.0 x yield strength to maintain piping integrity during a blast and there after the piping would in all probability have to be replaced.
Generally speaking, the fire case is a nebulus process/HSE condition in the line list and more often than not applies to the whole piping system. No account is taken of heat loss in the pipe regardless of where the fire originates or the actual lenght of the piping, and of course there is nothing in the stress design manuals about replacing pipe after a fire, so one has to assume that the system is expected to operate normally after as fire and design accordingly .
The only difference I see is in the nozzle allowables; if the fire case overloads your nozzles then perhaps the vendor will have to agree to increase the allowable nozzle loads due to the short term?? nature of the loading.


Regards
A

Hello,

A few thoughts:

I think that DESIGN temperature and DESIGN pressure are well defined by the Codes. Operating temperatures and pressures are usually lower than design temperatures and pressures as there will be "normal variations".

Some refining companies require a system inspection if an "upset" occurs. We design for most expected upsets (PRV's lift, seismic events, severe lightning strikes, etc.). These are things that should be included in the design conditions, but an "after event" inspection is due-diligence. This inspection should be covered by a "check-list" and sign-off sheets (especially if there are category M piping services involved). We would also want to take a good look at our instrumentation and control systems for damage from stray electrical current.

Sometimes there are more severe "off-design upsets" that ALWAYS call for a system shutdown and inspection (e.g., floods, 50 year wind, 50 year seismic events, potential damage from nearby blasts). Most of these events will force an emergency shut-down and it will be necessary to do an orderly inspection to assure there has been no damage (or, to derate if minor damage is found). As these things are not normally included in the scope of design, thorough inspections are needed to assure fitness for continued service.

All refining companies have a written procedure for what must be done after a fire. These procedures generally assume that the pressure piping and pressure vessels are "faulted" (damaged) until they are proven otherwise (e.g., we usually do a complete insulation off inspection and this includes hardness testing to look for temperature damage). Also, piping system levels will be checked to detect sagging due to over-temperature, and any relief valves that were near the fire will be removed and serviced. A fire is similar to sabotage in that it is not something that is included in the design conditions. A company may put into place some "safeguarding" to mitigate the effect of a fire or to limit or quickly control a fire but fire is certainly NOT a design condition.

Engineering is a science. When engineers include accidental fires in the scope of design conditions, maybe they should also drive main battle tanks to work.


Regards, John.

wow much thought;
what runs thru my mind as I read these notes is:
seismic and wind are more nearly inertial sustained conditions in that they must satisfy the laws of equilibrium, i.e., more nearly consistent with 'primary' loads.

Whereas displacement strains due to radiant temperature and external energy from a fire are more nearly self-limiting, i.e., consistent with 'secondary' loads; likewise I wouldn't consider a fire situation an 'occassional' condition as the occasional conditions mentioned (wind/earthquake) are inertial, i.e., primary, rather than self-limiting; occassional loads are upset sustained conditions. Upset strain-limited conditions are handled differently (as described below).

Therefore, the permanent deformation incurred by excessive internal or external temperature and radiant energy is of concern only as far as whether 1) the piping, or structure or attached equipment can handle same and 2) the deformation has not caused wall thinning, or flange warping, or gasket warping (or melting) such that the pressure boundary is compromised; that's to say the self-limiting strains are of concern only as they adversely affect the primary stresses and pressure containment.

The comment that flange leakage or gasket damage is likely to happen is probably the best bet.

But one must also remember that Displacement Stress Range is just that "A RANGE"; one cycle to some plastically deformed state after a fire is NOT a range; (see B31.3 para. 319.2.1(c) 2nd paragraph) that's like opening a paper clip; the clip remains open in its plastic state ... not functional but not failed (except by code standards). B31.3 for example pretty much assumes that piping displacement stresses (based on these strains) are going to exceed SY at temperature, and as long as the strain commensurate with SE is less than 2*SY at temperature, the the piping will shake down to elastic action.

So, it's possible that if the gaskets are replaced and all inspection care is taken to make sure that the supports analyzed in the analysis are indeed still there in real life (they've no melted in the fire), then SUBSEQUENT operating cycles will be at a new 'datum' but will generate a stress range that would be the same as before the fire ... except with a new initial position. Now if that new initial position is acceptable as unsightly as it might be, then the piping is functional.

Most 'owners' don't like unsightly piping and would be more likely to cut it out and replace unless there was compelling 'numbers' to prove the piping and structure were still adequate.

This would be accomplished thru B31.3 eq. (1d) where an equivalent number of cycles could be generated. I'd accept a one cycle stress state of 'fire-SE' above the oper SE (I know the code says the oper SE must be the highest; I've yet to have anyone tell me why that must be, and there's no reason some excursions couldn't be more severe than the most severe coincident set of pressure and temperature).

so you'd have a (fire-SE/oper-SE)^5 * 1 cycle;
add this to the other stress states' ratios to the 5th power ... each times the number of cycles they are anticipated to 'see' and summ those cycles; if the cycles are fewer than 7000, then a full fatigue cycle factor of 1.0 is still warranted for the 'ugly' piping; as Neq exceeds 7000, the 'f' reduces; as long as 'f' is not reduced more than the 'kitty' between SE and SA, then again, the 'ugly' piping is acceptable.

This same type process could be performed considering cummulative damage (basically the inverse of Neq).

Again the most important aspects to remember are
1) gasket integrity,
2) unacceptable deformation causing impingement, flange leakage, equipment alignment issues,
3) support structural integrity,
4) other large deflection issues (where sine(theta) is not assume to equal theta (in radians) ) if piping has elongated or deformed where bending or sag is more than a couple of degrees ... (you'll get results in such cases, but they'll not represent reality!)
5) and whether the plant manager can stand to look at 'ugly' piping; my guess is the manager 'can't'; as such I'd plan on 1 cycle (like aircraft brake-lockup scenario) with mandatory replacement.

Hot Topic... a good synopsis mr. Breen... I guess all I might add is that if one assumes that a fire is a "contained event" (in other words localized) one may want to protect the piping systems and vessels from the overpressure that could be caused by that event. This could be done with minimal effort as opposed to designing the system and its pipe supporting elements for exposure to an elevated fire temperature.

Hi,
I agree with most what John and Don are saying and it all makes perfect engineering sense if we live in an ideal world where we are dealing with piping subject to real fire conditions and are given logical boundary conditions and manhours to work with, but as I said before we are dealing with nebulus design/upset conditions that dont make sense such as 350C for a fire case, well I dont think CS is in the creep range at 350C so its pointless discussing plasticity, gaskets failure etc, you have to work with what you are given and what you are given is 350C. So what do we do.. ignore the 'flex' temperature and just work with the maximum 'operating' temp, I dont think so, unless of course we want to get fired, and all the while those iso's are piling up waiting to be signed off, no time for long deep discussions with the process engineer, just get the work out...only make sure its safe, and thats where I am coming from. Oh another point to bare in mind is we are talking about front end engineering here.

Regards
A

Hi Aaron...

Use Max Operating temperature UNLESS your contract with the owner specifies that your piping systems must be designed for a fire enveloped condition.

I won't forget years ago seeing 700F listed on a potable water line going to a safety shower! When I asked the process folks about it the response was "thats the fire case temp" when I pointed out to them that when the area was bathed in fire it was doubtful that anybody would need the safety shower the response was gee are you sure!

Hi Arron,

The basis (starting point) of any piping system design is a Design Specification that has been approved by the "Owner". Without that basis the design organization and the owner are setting themselves up for a "finished" design that satisfies neither of them. There is no room for nebulosity in the D-Spec - it protects both sides from the "evil" open ended contract.

If the owner wants the postulated fire event to be the "design condition" then it should be accommodated by the D-Spec. The temperature and pressure associated with the fire condition would become the “design T/P". The word fire need not appear (that word, used in a design document predisposes the owner to be liable for all sorts of things under the law) as it would be enough to simply state the design temperature and the design pressure in the D-Spec. It is the owner's call. Once the design T/P has been officially documented in the D-Spec it becomes a matter of designing to the Code and by precedence to the municipal, State, Province...Building Code. The owner and the design organization would be well served to protect themselves. THAT is the real world. That world is not an ideal world. To do business any other way is a "**** shoot" that represents risks that neither side can afford.

I can't believe I have (after all my posts) inadvertently evoked the "cyber censor" in the description of a game of dice (I guess that puts me into John Luf's league now). Well, that at least gave me a chuckle.

Regards, John.

HiJohn,

Your points are valid points but I don't think they are addressing the situation at hand. I will put it another way:
*first we are dealing with front end engineering *the piping stress design specification has nothing on how to treat fire case (which is quite normal in this industry).
*the process line list shows a normal operating temperature 200C maximum operating temperature(design) of 235C and a flex (fire) temperature of 350C, so we have to assume that this means that the owner specifies that our piping system must be designed for a fire condition
*the pipe is not in the creep range at the flex temp
*the question is, which temperature shall we use to do our expansion stress analysis and at what temperature will we get the maximum terminal loadS? John Luf says use maximum operating temp
so does that mean we ignore the flex temperature???
I look forward to your comments

Regards
A

Hi Aaron,

I don't think you ignore the fire condition temperature, I think you consider the pipe to be in a faulted condition (subject to mandatory examination - per fire recovery procedure) if there is a fire or if for any other reason the pipe temperature goes to 350 C. And as Don says, there are a lot of things that must be examined. For design, I would go with the design temperature of 235 C.

I have been in the industry for over 40 years and I still do not consider a fire "quite nornal" (I have done fire recovery for 4 or 5 companies). Yes, fires happen but (thankfully) they happen less often due to the industry's commitment to safety. Don't build the church for the Easter crowd.

I have some photos of an NPS 6 steam system (mostly horizontal) after an upstream desuperheater failure and it looks like home made pasta hanging (drooping) out to dry. Sometimes hot catalyst gets into piping that was not designed for it too. In these cases (no fire) the piping must be examined and its temperature embrittlement must be quantified (albeit, if it is carbon steel pipe (relatively inexpensive) the owner may opt to replace it as NDE adds nothing to the value of the pipe).

There will be some report that documents the design and that is the place to declare the design parameters and state the assumptions. One of the assumptions would be that if there is any over pressure or over temperature event (greater than "variations") the system must have an orderly and well documented examinine per a specific recovery procedure.

Regards, John.

Hi Aaron,

I feel your pain and confusion…

The first thing you should do is to inquire what the owner means when they use the word “flex (fire) temperature” in the line list it does seem to indicate that this is one of the many possible temperatures they wish to use for flexibility analysis.

This temperature is far below the actual temperature that meat would be exposed to in the case of direct fire impingement so I suspect that it’s based on the concept as follows…

In an operating unit you have vessels tied together with piping… if a localized fire develops around a vessel the vessel and its contents will heat up. As the contents heat they will rise in pressure and temperature eventually opening up and then through some PSV’s etc. Then the contents will boil off through the vent(s). During this time the fire would hopefully be local to the vessel and not impinge on the pipe. However the owner wants the piping system to be designed for handling fluid based on this. (Assuming I have read between the lines correctly)

I have designed on this type of basis but when doing so you must examine all the different temperatures for displacement stresses as well as examine all the sustained loads and occasional loads based upon the various temperatures (This is required by B31.3)

Now when I have been involved in this type of basis it was required by the client to have retention supports for springs and also required fireproofing of support elements. Another point of caution I advise is to make sure any cans are based upon the Max Op Temp and then check the cans for max travel.

Read the contract and ask questions talk to the process engineers and ask them what the flex temp is based upon… this is very important before the start of work usually people always overlook this step, instead they rush into doing the work, work that may require being redone later.

I have never seen conditions for a fire case. I have, however, had to analyze for cases due to the loss of equipment, often called "loss of cooling", possibly from fire. As mentioned previously, if one piece of equipment is lost or operating outside the intended parameters in the system, it can have an effect on everything downstream.

As for analyzing this case, I typically design the system (nozzle loads and code stresses) using the Design P/T on the line list, then check the system stresses using the "flex" conditions on the line list. However, as I type this, I am wondering which Sh will be used to check primary stresses. Should it be the Sh for the design condition? If all cases are run in one file with all temperatures, I think CAESAR will use the Sh from the highest temperature. Is this what the code intends? Should the "fire" case be run in a seperate file in order to use the Sh for the design conditions?

About the only system I've ever heard being designed for direct fire contact is an FRP firemain on offshore platforms. And that was not a physical case analyzed by stress, but a requirement for possible FRP Vendors to satisfy in order to be considered.

From what I read, the 350°C case is more likely a "loss of cooling" scenario.

Richard,

B31.3 has rendered numerous interps on which temperature shall be used for sustained or indeed occasional stresses.

All of them shall be evaluated... in the state of stress as they are actually loaded in other words a hot sustained case with inactive supports removed from the boundary conditions.

Hope this helps clarify a confusing issue!

Hi John,
All your assuptions are correct!
I have already discussed this matter with the process engineer and know the cause and source of the fire but dont agree with some of his assumptions relating to the flex (fire)temperature distribution in the piping system, however I have to accept his figures.
The fire is expected to be localised at a vessel and heats up the contents of the piping all down the pipe rack. We have springs in this system but away from the source of the fire, and yes, I have already considered the spring over travel due to the flex temp.
All cases have been considered, but only by considering the flex temperature (350C) as the maximum operating temperature. I am still not convinced that I should analyse the system to the lower temperature which is the process design temperature of 235C.
Now, I am coming back to my original question, do I ignore the flex temperature?
I am beginning to feel as if I am flogging a dead horse here.
Incidentally for what its worth, this is not the first time I have come across a fire case in the line list, its quite common in front end engineering, perhaps due to a greater involvement of the HSE.

Regards
A

"Now, I am coming back to my original question, do I ignore the flex temperature?"

Nope... you should design for this temp. If you keep digging you will find somewhere in the contract requirements to use this "fire case" temp for design it will be an owner requirement... but keep digging....

Nothing ventured nothing gained... (B. Franklin)

Hi John,
Thanks for responding, I feel I have gone full circle with this topic and now feel relatively comfortable with my approach and solution to the problem, and of course I respect all your comments.
I think this type of dialouge is invalueable as it opens many other windows of thought, such as those generated by yourself, John Breen and many others, thanks and keep up the good work.

Regards
A

Dear Aaron & Captain Kenny,

I admire the self confidence of both of you.

But, what is the harm of thinking out of the box, thinking flexibly like a small child!

If the owner's requirement asks for considering fire case as per API RP 520 or any other applicable code, it may be possible to design fire case as a single occurance secondary load 'faulted' event - like tank settlement or SSE SAM(seismic anchor displacement) ? This may serve the purpose, at the same time not becoming too costly as an alternative.

While remaining focussed within our contractual obligations we must not forget our alliances with owners - they pay us for their SAFETY within their budget, we must stay true to our roots to safeguard our brand image. We have to keep on looking hard at all the options.

regards,

sam