Hi
I 've got a system with three PRV's on a manifold with a design temperature of 218C & operating temp of 178C. The PRVs are handling a vapour from the top of a tower and are situated on the tower top platform and are discharging directly to atmosphere and the pipework expands from 8" at the discharge quickly up to 18".
In the absence of more informed process information, what would be a reasonable temperature to use to model the discharge pipework?
Operating temperature seems a bit conservative, given the valves should not be relieving for that long and are open to being 'cooled by the breeze' as someone put it; also the rapid expansion of the gases will probably result in a drop in gas temperature, so I'm considering using 80C as a half way between ambient and operating.
With such a small valve [6" x 8"] and large tailpipe I'm struggling to get the moments down on the valve itself as any expansion in the inlet leg is not matched by that in the cold trunnion under the discharge elbow. The differential expansion isn't much, but it is generating very large moments.
Any thoughts on either of these issues anyone?

I don't know what piping code you must comply to, but in ASME B31.3, for temperatures 65C and above your allowed to use 95% of the fluid temperature for uninsulated pipe.

Of course B31.3 allows you to use better information if you have it. I don't think taking the average between the ambient temp. and the operating temp. qualifies as a heat transfer calculation or better information.

Kenny,

Are you saying that each of the PRVs discharge into individual 18” headers each of which blow to atmosphere? If so, it seems like an excessive line size increase, particularly since the discharge length is quite short (I assume the discharge pipes turn vertically up immediately after the PRVs and the open end is approx 10’ above the platform). You might question the process engr to see if you can get the size reduced.

Also, I’m not clear on why the temperature in the discharge makes any difference to the moment in the valve. If the vertical section of the discharge pipe is growing away from the trunnion (I take it we’re talking about a base-elbow type support here) then the pipe is simply growing away from the elbow. The moment (again, making the assumption that the moment is caused by the deadweight of the discharge pipe) should remain the same no matter what the discharge temperature is. It would seem, if all assumptions are correct, that the problem occurs when the trunnion lifts off due to the temperature change in the valve inlet piping and valve body.

Have you considered using a separate vent pipe arrangement? I have in mind the type of arrangement pictured in B31.1 Fig.II-7-1. This piping arrangement is more common in power plants than petrochem facilities, but its advantage is that the only deadweight on the valve is a short section of discharge-size diameter piping. It will, however, require a separate structure to support the vent pipe but that should be possible even on a top-head platform. A second advantage of this arrangement is that wind loads on the vent pipe are absorbed by the structure rather than the valve body.

Regards,

El Gringo
The expansion up to 18" does seem excessive to me, but then I've not sized it, and if that is what process need, well....
The moments are coming from two places.
1) the vertical growth to the valve on the inlet side, either causes the trunnion to lift off and then the discharge reaction has nothing to restrain it [or if I tie the truunion down, the expansion causes the moment]. This is the biggest cause of moment. It looks like if I 'cold push' the trunnion up a few mm, I can minimise the moments caused by this,, but I do realise that this is never going to be anything other than educated guesswork on site to install.
2) Depending on the discharge temperature used the long horizontal section of 8" RFWN flange, two reducers [to get from 8" to 18"] and a 18" LRWE, does tend to grow a bit. Depending on how the trunnion is tied down, this can have an effect.
My original question was that if we assume the discharge piping is cold before the valve blows off, it will take time to reach full [or 95%] operating temperature. If we assume the relief is a short term effect and there is no joule-thompson cooling effect, the discharge piping will reach a temperature somewhere between ambient and operating. What would be a reasonable assumption here.
As for the temperature of the trunnion...well that is yet another question for another post I think!!

I would suggest:

1. Give the process poeple a slap and tell them to stop being silly.
2. Try to get the support before you change dia.
(ie after the elbow)
3. Guide the trunnion only (no anchor)
4. Possible change the Sched of the discharge pipe to a thinner wall to become more flex
5. Possible increase the Sched of the Valve inlet pipe to a thicker sch to become stiffer
6. Slap the process engineer again.
7. assume that on the first opening the time is sufficient to heat up the trunnion.( although cold on first reaction, and that the valve can again open (or chatter) with the trunnion hot.

good luck

You can always trust the Irish to talk sense.
Another consideration is that the PED requires us to:-
"apply appropriate protection measures against hazards which cannot be eliminated,
- where appropriate, inform users of residual hazards and indicate whether it is necessary to take appropriate special measures..."

I guess slapping the Process guys meets both of those requirements

Kenny,
I would be intrested to see how your firm is approaching stress analysis with respect to PED.

For us, we still use Caesar/B31.3

B31.3 / caesar is pretty much the norm, as the PED doesn't say much about the matter. Of course you do need to make some allowances [impacts, over-pressure allowances etc]. I don't think any of the NoBos have any problems with the B31 codes.
I personally have a lot of time for the new Eurocodes. 13445 [vessels] has excellent analytical methods for doing local loads on shells without using graphs.
13480 [piping] is pretty much a copy of B31.3, and although I have not spent much time with it, it does include a published method for doing trunnions and clips [I know these are dear to your heart!].
The Eurocodes are 'harmonised' which largely means compliance with them means compliance with the relevant sections of the PED.
What do JE do?

They used to be a document from eemu covering compliance via b31.3 but this has now been withdrawn.

We don't use EN13480.
the idea i gather, is a project note, covering the use of B31.3 for PED.

I don't think we do it properly, but i dont know what properly is, most Eng's i talk to have no opinion.

I would hope to do one or the other,i can't imagine being an expert at both.

Clarity would be appreciated.

Kenny,

First, the temperature issue: In the past I’ve modeled PSV discharges at ambient temp but you’ve got me thinking; what is a realistic temp of the line after the valve opens? Clearly it could be anything from ambient—assuming a very short burst of the valve—and it could be as high as design temp. (I have seen relief valves in steam systems remain open for more than an hour, so it’s possible to at least reach operating temp. And if we assume that your process PSV may open during a major upset condition you could have a situation somewhat higher than operating.) To state the obvious, we cannot predict what a PSV will do during its lifetime. Now, one approach I’ve taken with unpredictable temps is to model a range of temps: Say T1=lowest ambient, T4=design, and T2 and T3 are intermediate temps. (I’ve probably reverted to something close to your initial “half way” idea, but my approach is at least an angstrom or two closer to science.) That ain’t gonna help you, however, so let’s look at some possible solutions to the moment issue.

I had a very similar situation to yours a couple of years ago. During that frustrating period when it seems that the thing is never going to work I considered using a spring under the elbow with a stop to prevent downward movement. Realizing that I might one day be endangering passing aircraft I quickly abandoned the idea. My final solution was to increase the distance between the PSV and the elbow (which in fact moved the ell to a more natural support point) and used the vent pipe approach I discussed earlier. The vent pipe had to extend up about 50 feet so I anchored it near the bottom and guided it at intervals thereafter. I also did some careful guiding of the trunnion to prevent it going anywhere I didn’t want it to. I was able to inspect the work prior to start up and even witnessed a discharge test during operation. All went well and to my knowledge it’s still a happy relief valve.

Other than that I can only agree with TJN’s idea: Put the elbow directly on the flange with the reducers in the vertical run. I’m sure that’s going to cause another wince from process but I doubt they can prove it’s really detrimental. This does however give you a weaker structure so you may have to add a guide to the vertical pipe run.

TJN’s consideration of pipe wall thickness is also valid (process sometimes forgets to change spec after the PSV along with including a hydrotest requirement in the line list) but I wouldn’t agree with taking any credit for the trunnion heating up—our assumptions of the pipe temp are approaching voodoo science even before we make guesses as to what happens to pipe attachments.

A couple of final thoughts: Unless it contradicts the client spec there shouldn’t be any insulation after the PSV. Sometimes that gets forgotten on the P&ID along with the spec break. Personnel protection perhaps. I doubt it will ever be necessary but since we’re in the business of eliminating risk I’d leave that to the client or process engr’s discretion. Also, I know you’re going to look at reaction forces but don’t forget the wind affects. A good Scottish gale at the top of a tower can cause an awful lot of movement on your 18” lightning rod!

Regards,

My point re temperature on the trunnion is not to gain credit, but to ensure a realistic model, which is probably worse.


I think you should Never,Ever put springs on a turbulent system such as a relief line.
I dont like springs anyway so i don't usually use them.

I don't like springs either. I just have a lot of shares in Grinnell so I just stick them everywhere. I even put springs on lines operating at 70F and they work perfectly!

Mr Kenny,

Your problem may reduce somewhat by insulating the whole trunnion( i.e. You take the temperature of main line in the trunnion also)and that will decrease differencial expanssion if you took support at same elevation at Inlet Leg as taken in Trunnion)

Best of Luck

Kenny,

As an answer to your original question for the design temperature of safety valve discharge piping, AS4041 Clause 3.9.8.3 gives:

T ( deg C ) = [ Ts - 41 ] / 1.15

where Ts is the temperature ( deg C ) of steam at the safety valve body inlet.

In case of long discharge piping I would suggest to use Open Discharge System given by "ASME B31.1 Appx II-Rules for the design of safety valve installations" by using a expansion chamber close to the discharge elbow to eliminate additional forces of the discharge piping and the thermal expansion.
However, this requires teh carefully selected expansion chamber and sealing system in the bottle. The main pipe might be moving relatively against the discharge pipe. The chamber should allow this movement.
So, you can isolate the discharge piping and solve and support it seperately. In this application the selection of discharge pipe diameter beyond the expansion chamber is important.
I believe you will find all the details in Appx II.

Hope it helps.

I. Demir