Dear all,

I am just seeking for opinions on the following scenario.

If we have an equipment with liquid content on floor level and its PSV is mounted ~5 or 10m in the air which is also liquid filled up to the suction. If the equipment is on fire, and the PSV is sized for this case. What would we have in the first instant? I understand heat input to the equipment would vaporize the liquid and the rate of vaporization governs the relief rate. But if only the equipment is on fire the vapor travels all the way up and lifts the PSV. Because the PSV is liquid filled, we will first have two phase relief?

I am wondering if two phase sizing or vapor sizing should be used? Normally if the PSV is mounted on the equipment and its on fire, only wetted surface area needs to be taken into account and PSV is sized under API 521 5.15. However, for this case, I tend to think two phase sizing is more reasonable.

Your views?

Thanks

This has been discussed briefly in "Psv Fire Sizing - Dense Phase, Relieving P above Cricondenbar (click here)". It seem Wendy is not convinced.

I think Wendy is looking second voice. Phil, Mr. Montemayor, Latexman, CheJedi... PLEASE SHARE YOUR OPINION....

Personally i would like to hear second opinion.

I think this is a different issue.

Anyway for a liquid filled vessel, a fire case does not justify two phase relief. API 521 para 5.15.3.3.

Also here is an excellent reference: http://www.fauske.com/Download/Chemical/Pr...y/Vol12-No1.pdf

Regards
rajiv

Rajiv,
Thanks for your inputs.

I have some opinions on the understanding the statement in API 521 para 5.15.3.3 and justification on the its application.



Post #4 in http://www.cheresources.com/invision/index...art=#entry16468 has briefly discussed potential 2 phase flow in liquid fully / partially filled vessel.



Strictly interpret from the text, looks like two phase flow relief is not normally consider in fire case BUT not absolutely discard the two phase relief. One of the precondition is "non-foamy" system. This needs detail analysis if the mixtures is foamy and API did not provide any guideline differentiate and justify a foamy and non-foamy system. The energy and effort put to justify foamy / non-foamy is tremendous...

5.15.3.3. has admitted that liquid swelling (no dicussion on liquid entrainment due to KH effect) is possible and it is occur at interrim period, however has stated "...mixed-phase condition is usually neglected during sizing and selecting of the pressure-relief device..." and "Two-phase relief-device sizing is not normally required for the fire case..." somehow NOT 100% discard mixed phase relief. The statement invites some serious level of engineering judgement before we can discard two phase, (or mixed phase to be excat) relief.

Joe,

I agree with your cautions. What I quoted is only for non-foamy liquids. For foamy liquids or liquid swell, DIERS manual should be consulted.

I am talking about normal hydrocarbons, and is Fuske associates articles 'which was hyperlinked. Interesting to note that he prrdicts that the pressure peak will not cross 21% accumulation.

Incidentally I am doing some depressuring study of liquid filled systems, and HYSYS is able to simulate liquid swell too. It is interesting to note that HYSYS does predict pressure peaks, but none of the peaks are worrying.

Thanks for your interest and response.

Regards
Rajiv

Another show case that what if API states that two phase flow is not so important while considering vessel on fire, all should correct API and start applying two phase flow because Boss wants so. Change API or resign.

Wendy/Joe/Rajiv:

I am surprised that logical thinking about a presumed complex case is being carried on when the original post is lacking in detail, logic, and a full description and explanation of what the complete scenario is composed of. I am not dwelling on semantics or pure English language grammar. I’m referring to the original post’s bad communication and lack of an accurate engineering description of what is happening physically. Please refer to my attached Excel sketches to fully (& quickly) understand precisely what I am referring to.

No mention is made of precisely where – on the vessel – the PSV nozzle is located and how the piping is routed. I really have to mention what I believe everyone fully knows already: It makes a big difference on where the PSV nozzle is located and how it is connected on the vessel. For example, how can we discuss a vertical, dead-ended pipe that is presumed to be 100% full of liquid. This never exists in the real world during a full process run. It doesn’t even exist in our houses, where any experienced plumber will quickly tell you that the water piping in the house has to have an “air-trap” (nothing more than a vertical pipe riser that is dead-ended with a pipe cap and serves to trap and collect air) or two at the highest points in the piping network in order to absorb pressure surges and water hammer in the system. It is common knowledge that any vessel or pipe in a liquid system will eventually collect non-condensable gas or vapor at any convenient, dead-headed, high-spot location in the system – especially if it is normally static fluid, just exactly as the fluid in the suction to a PSV is.

Additionally, it also makes a big difference if the PSV is connected to the top of the vessel or to the side of the vessel. How can a vertical standpipe be filled with liquid without having a vent at the top? The answer is: it can’t – period. It may be an academic assumption, but it is a real-world impossibility without additional valving and special operations – which no one practices in real industrial applications. This is probably why almost every vessel PSV I have seen, designed, and installed has been on the top of the vessel’s liquid level to ensure that vapor pressure is relieved – not liquid.

There is no such thing as a generic 2-phase flow. There are approximately seven (7) type of 2-phase flows conventionally identified (there may be more – but we haven’t identified them yet) and it is a known fact that it is virtually impossible to identify which one is taking place with a reasonable amount of accuracy – especially any vertically flowing mixture. We can only make academic and logical guesses as to what may happen when a vapor and a liquid fluid join together and form a common flow stream under defined, physical configurations. This is a generically tough and complex problem to solve with any degree of “accuracy” as to type of flow and resulting pressure drop. Work is still continuing in this field and much more is still required. Fauske and Associates are renowned for their work and correlations on 2-phase flow. They are probably the acknowledged, top experts in this field. However, even Fauske can’t begin to answer Wendy’s question. Some of the first questions he (or anyone else) would ask are:

1. Where, exactly, is the PSV connection nozzle located on the vessel and how is it configured?
2. How is it possible to ensure that the vertical PSV standpipe is going to be 100% liquid full at the time of PSV activation?
3. What are the fluid(s) in question and what are their chemical-physical characteristics? Rajiv brings up a very good point about “foamy” liquids. Are they?

Before spending much more time and effort on what could be an important and interesting subject, I suggest some effort be made to correctly, accurately describe the scenario in question – in engineering detail. For this type of problem, a sketch is not only worth a thousand words, it is a prerequisite towards fully explaining what is being communicated.

Wendy asks: “What would we have in the first instant? I understand heat input to the equipment would vaporize the liquid and the rate of vaporization governs the relief rate”.

My comment is that it is more than obvious that the initial fluid into and out of the PSV will be gas or vapor – liquid may or may not follow, mixed or unmixed. It fully depends on the configuration and location of the nozzle plus the fluid’s properties. Additionally, it is popularly believed – especially by students – that a liquid-full vessel will “boil” its sub-cooled liquid contents when subjected to sufficient heat input. This is not true in the sequence of important engineering events. I see this error all too often and it makes for propagating wrong deductions about physical laws. A contained, pure sub-cooled liquid system will gain sensible heat first,, upon being subjected to an external heat source like a pool fire. This sensible heat will cause an expansion in the sub-cooled liquid inventory with an increase in hydraulic pressure. Depending on the liquid’s coefficient of expansion and on the availability of a static gas chamber within the vessel, the hydraulic pressure will continue to expand until it reaches the PSV set pressure and is relieved by the PSV. If the external heat continues, the vessel liquid contents also continue to sensibly heat up until reaching the saturated conditions corresponding to the pressure inside the vessel. At this point, latent heat is absorbed by the liquid and converts a proportionate quantity of saturated vapor (at the internal vessel pressure – maybe the PSV’s set point). It is only at this stage in the sequence of events that the vessel begins to generate continuous vapor (“boils”) that quickly increases the vessels internal vapor pressure. This is where the PSV’s design capacity has to be ready to “kick-in” and vent the continuously generated vapor without allowing it to accumulate and exceed the MAWP. This PSV inlet relief stream may involve single-phase vapor or, in the worse case, a 2-phase mixture. And this is the critical stage at which I think we should be focusing, if we are concerned with the PSV’s capacity. Any reasonably experienced engineer should know that the relief of hydraulic pressure caused by contained liquid does not require a significant amount of PSV relief capacity. A ½” or ¾” expansion relief valve should be able to take care of hydraulic pressure build up. It is the impending latent heat addition to the liquid that causes the subsequent rapid pressure buildup that must be relieved safely.

I apologize for this rather long post; I was asked for comments and the comments that I have require a detailed explanation in order to clearly identify what I mean by the shortcomings of the original post. Any vessel relief query is an important query that should be taken seriously – just as Joe and Rajiv always do. That’s why I believe in precise and accurate description of the scenario and the equipment involved. Our members can resolve any PSV relief questions – but they require a precise and accurate description to do it correctly.
Click to view attachment

Art,

Appreciate your time invested and the reply to the query. I agree with your approach, requiring a detailed description of the the problem. As one can see in the forum, the details of a problem are seldom given by the originator, so the replies are correspondingly sketchy.

I wanted to give a quick response to the stated problem, as I understood it, and which is quite common. In fact, I had replied to a similair query in a different forum recently.

In my enthusiasm I can add more replies to Wendy's rejoinder, but I think I will refrain, pending a complete definition of the problem!

Thanks to Joe, for contributing too.

Regards
Rajiv

Thanks Art, Joe and Rajiv for your excellent and detailed post. I really appreciate your time spent and I can see this leads to some very good discussions and definitely help me and other readers understand PSV sizing more.

I have to then apologize for my poor post in the first place. As Art pointed out I did not make the query clear and it is lack of an accurate engineering description. I'll try to make it up here (I have been away and just accessed the forum)

What we have in the first deck is a heater with ~85% ethylene glycol solution being the process fluid and it is heated by hot oil. The outlet piping runs up to the third deck and goes to a process vessel. The PSV suction line runs off the outlet piping and it is mounted up high ie. 10m above first deck. There is a vent valve before the PSV suction nozzle and there is a bypass line. The PSV suction piping can be liquid filled if we open the bypass valve or the vent valve. Please refer to the sketch attached. I haven't done the sketch properly but hopefully it gives enough indication.

One of the PSV sizing case is the heat exchanger on fire with/without its associated piping. The first sizing was done based on the vapour generation rate from the fire input but then I had second thoughts maybe two phase flow should be applied.

Thanks

Wendy,

I first tried to find out about the properties of Ethylene Glycol. I presume this MEG? Anyway I found the following in Wikipaedia:

Ethylene glycol can begin to breakdown at 230° – 250°F (110° – 121°C). Note that breakdown can occur when the system bulk (average) temperature is below these limits because surface temperatures in heat exchangers and boilers can be locally well above these temperatures.

However, I could not corraborate this with other text books I have. Could you first have a look in your references? Knowing how the fluid behaves at higher temperatures could lead us to the correct path. I tried to find out if the fluid has a foaming tendency or not, but was unable to.

Regards
Rajiv

One of the operation problem with MEG is present of HC in MEG solution and severe foaming in MEG regeneration column. If HC present in MEG solution, there is the tendency of foaming.

hi everybody,
Thanks all about the information above.

Now, I am calculating a PSV which is installed on the piping ( no any equipement on that), that mean this psv for object decreasing the hydraulic expansion by thermal. for this case I have two sources of heat: one from solar radiation and one from ambient heat gain.
my question is that when I use the heat source from solar radiation for calculation? and when I use the ambient heat gain?

Sorry for my english is not good.
thanks,

Linda,
The answer to your question is that you use both. You always calculate all possible cases to reach the point where you can identify the GOVERNING case. After that, you typically need only worry about that single case (although there are some instances where multiple cases may be important).

To All,
I like Art's approach and clear presentation of how one should plumb up your PSV in liquid service. I especially like (and favor) locating the PSV inlet at the vessel's highpoint. If this advise is followed, even accepting Wendy's contention that one might somehow get liquid right up to the PSV, what are we worried about? There may be a short period of time where the liquid is absorbing heat and expanding. Do the calcs and you'll see that the volumetric change of cool liquid going to warm liquid is very small. Furthermore, the mass and heat capacity will be large for a liquid filled system. I believe it is safe to totally neglect that short period of time.

Once at full relieving conditions, further heat input generates vapors (at much larger volumetric flowrates). If the PSV inlet is sitting in what would now be the vapor space, then only vapors will be relieved. If the PSV inlet is submerged, then the vapors generated will "push out" or displace an equal volume of liquid. Essentially, at all times the fluid flowing into the PSV will be either vapor or liquid (i.e. not both as long as we're not worried about foaming situations).

So far, I believe that the only thing I am guilty of is neglecting the short transition period that may exist before starting vapor flow. But now we can see why it is best to avoid submerging the PSV inlet line. If we continue and force hot (saturated) high pressure liquid into the PSV inlet, it will flash as it passes through the PSV. The internal and outlet volumetric flowrates will GREATLY exceed the inlet flowrate. This I believe represents the proper description of two phase flow. This condition may severely constrain the system's ability to relieve pressure and may result in a (much) larger PSV and outlet piping than would otherwise be the case. This is why I heartily endorse Art's contention that it truly,
"It makes a big difference on where the PSV nozzle is located and how it is connected on the vessel"
I wish I had more knowledge of how the codes handle these situations, but I don't often do PSV calcs. Perhaps someone more knowledgable could address this deficiency.
Doug

thanks djack77494,

Can you explain for me how do I estimate the Ambient Heat gain? Because I don't understand clearly about cause of Ambient Heat gain yet.

Note also that:

• the static liquid head created by the 5 - 10m elevation of the PSV could be significant and require the set pressure to be appropriately less than the vessel design pressure or MAWP
• the long inlet pipe will need to be appropriately sized - dP < 3% of set pressure
• the pressure in the vessel will change as the contents go from 100% liquid to liquid and vapour and the vapour flow (and associated dP in the inlet pipe) and allowable over-pressure develop.

linda,
Ambient heat gain merely refers to the fact that the ambient temperature may be higher than the contents of your system. When this is the case, heat will flow from the environment into your system. Search for heat gain or heat gain and you should be able to find a spreadsheet or at least the equations needed to calculate these values. It seems to me that I've seen one (I think right here) that was designed for estimating heat losses from a pipe. That would be ideal.
Good luck and sorry it took me so long to get back here,
Doug

with your reference to your comment “Where, exactly, is the PSV connection nozzle located on the vessel and how is it configured?”
I would like to highlight another problem regarding take off point of a PSV for high elevation columns. I have seen some Columns PSV takeoff point from columns top and some other columns the PSV takeoff point from the vapor return line to the condenser. Most of designer claims the benefit is piping cost. Where as I have seen both design exist in previous company designed by same designer and the difference was Deethaniser & debutanizer columns with almost same elevation. Please comment your opinion about this

With reference to following comments made by Mr. Art Montemayor regarding expansion relief valves ,
“Any reasonably experienced engineer should know that the relief of hydraulic pressure caused by contained liquid does not require a significant amount of PSV relief capacity. A ½” or ¾” expansion relief valve should be able to take care of hydraulic pressure build up”

I like to know about the thermal expansion PSV function. We got single thermal PSV in ethane reflux pump suction line as well as butane reflux pump suction line, whenever these PSV are taken for maintenance, we are isolating the particular pump and depressurizing. In addition we are de-engerizing the pump motor power supply too. We got two set of argument regarding the isolation procedure for this thermal PSV related equipment.
With above operation it has been treated the thermal PSV like normal fire PSV or the PSV on discharge of positive displacement pump which requires pump isolation. Note: Pump being isolated by block valve only not by spading and there is chance of getting pressurization due to block valve passing or vaporization holdup liquid.
In other way, Thermal PSV given to protect the pump during isolated the condition, which may subsequently buildup hydraulic pressure due to ambient conditions. So pump has to be kept on unblocked conditions to avoid any hydraulic pressure buildup which may get equalize with system pressure. This PSV must be on service when the pump kept isolated. So during PSV maintenance, pump must be kept on line.

I hope you will give us correct procedure to follow for thermal PSV.