PSV orifice area 5

[farzinchemical]

hi everybody

I have sized a PSV. the calculated area is very small (0.44 sqin). the PSV's on a similar existing installation are 11 sqin. can anyone give a hint?

 

[MortenA]

Your calculations are wrong?

Check if its in the calculation of orifice itself or in the relief cases that you have identified the difference lies.

Best regards

Morten

 

[roker]

advise your data for the two cases and will try to recalculate.

roker

 

[Montemayor]

farzinchemical:

Correctly sizing the orifice area for a PSV has little to do with the mundane mathematical exercise of working the equations - whichever ones you apply. However, the correct orifice size has EVERYTHING to do with the correct over-pressure scenario being identified and correctly calculated as to the capacity required of the PSV under relief conditions. The mathematics involved is elementary application of algebra at best.

I would not get very concerned about a degreed engineer doing the math correctly. What would concern me first is: did the same engineer correctly study, analyze, and identify the worse credible scenario? Now, that's real engineering applied where it should be. What is the design base for the existing installation, and how does this compare with your worse case scenario? If no calculations exist for the existing PSVs, I wouldn't give any credence or recognition to the existing PSVs as being correct in the first place. Secondly, if you have no calculations and case study of the worse case scenario for your PSV then the situation is even worse and you should not proceed any further until someone experienced and capable of doing the required, total design task is assigned to it.

If you are dealing with a thermal expansion scenario or application, I'm not surprised by a very small PSV orifice requirement. If you're dealing with a 2-phase, reactor excursion relief application then I've got a lot of problems accepting what you come up with without requiring a complete package of ALL the basic data and Scope of Work. You don't furnish either, so I can comment no further.


Art Montemayor
Spring, TX

 

[farzinchemical]

thankyou everybody,

I am surely not experienced enough. oherwise i would not ask for help.

let me give more details.

it is an oil-gas separator operating at 370 psig.
MAWP is 450 psig.
set press. 450 psig
allowed overpress. 10%
gas outflow at steady state 6.7 MMSCFD
oil inflow at steady state 30254 bbl/day
backpress. 8.5 psig (constant)

i have assumed that at operating contingencies, non-fire case, blocked flow condition will force the PSV to discharge 6.7 MMSCFD of gas.

is this a right assumption?

 

[Montemayor]

farzinchemical:

As I tried to emphasize in my previous post, this is the most important phase of sizing the correct capacity required by a PSV: the phase where you identify all the credible, possible scenarios and identify the worse of the lot. No one on this forum is in a position to evaluate all the over-pressure scenarios possible in your process. We don’t know if we have all the required basic data. We certainly don’t have the “As-Built” P&ID – which is the first important item to have in these cases. Therefore, we can’t verify for you that you have made the “right assumption”. This exercise is not about assuming anything. You should – and I’ll repeat it once more – identify the worse possible credible over-pressure scenario. And this doesn’t just necessarily involve “normal” operating conditions. It must cover ALL situations where it is possible to have the separator subjected to over-pressure – for whatever reason. Those are the recommended practices set forth in API 520 and just about every process industry world-wide.

If the blocked gas discharge case is the worse case that you have identified while listing (and quantifying) all the possible credible cases - then fine. You have probably arrived at the maximum required PSV capacity. But I can’t verify that for you. I don’t have your basic data or Scope of Work tools. You and your fellow workers are in the best position to make that decision.

I’ll be the first one to admit to you that this is a very mentally taxing exercise that requires a lot of thinking and rationalizing – as well as a lot of calculations and data retrieval. But it is the ONLY ACCEPTABLE and proven way to ensure a safe operation. It has to be done – and done diligently and accurately in all details. Otherwise, you will wind up calculating and installing an unsafe PSV – just the opposite of what you have been assigned to do. I hope you can understand now that the actual calculation phase of this project is the easiest. The formulas are simple and they are carried out by a computer with simple software. The hardest part – and the heart of the whole operation – is the identification of the worse case scenario. I hope that you have already arrived at this stage. If not, then you should go back, and with a large bottle of aspirins go over all the detailed logic in every possible over-pressure case that you and others can identify as credible and calculate the capacities for each. This is tough, grinding work, but it has to be done. People are counting on you to do a 100% accurate job – otherwise, their lives may be in future jeopardy.

Good luck.


Art Montemayor
Spring, TX

 

[jsummerfield]

Find the calculation for the similar existing installation. This might lead you to the original assumptions.

John

 

[farzinchemical]

thanks again everyone, especially meontemayor.

unfortunately there is no datasheet for the existing PSV's. the existing unit is 30 years old and the scenarios for PSV sizings is not known. all we have is what we can read on the non-asbuilt P&ID's. all we know is what has been written on the P&ID's. two parallel PSV's,6"Q8" having set pressures of 520 and 580 psig respectively.

I have searched a lot and as you have noted the heart of the calculations lies in determining the worst case scenarios. can anyone lead me to an example of such analysis. all you can find in API 520 is technical stuff written for professional process engineers who can analyse the different scenarios very well. can anyone lead me to an actual analysis of a sample case.

thanks again

 

[Montemayor]

farzinchemical:

Thank you for responding and advising us with the rest of the story. There are many other experienced engineers on this forum who, I am certain, share the empathy I have for your situation. I have led process teams in the past on projects to cure the very situation that you are describing: a plant where safety was secondary and the maintenance of process information and records was not kept up or had been discarded. You and others would be very alarmed if I were to reveal the names of these rather large companies. If you presently find yourself in a country where safety has been mandated as the responsibility of the operating plant and its management, you are in a very serious situation and one that must be resolved timely, correctly, and safely.

I also believe you share with me the realization that you must have “As-built” P&IDs, PFDs, specification sheets, equipment drawings, and instrumentation logic. If this is not the case (as you reveal), then you can resort to what I (and countless others) have done in the past: literally walk out the physical equipment site and verify the equipment, piping, instrumentation, controls, and process capacities. Take a large sketch pad with you and especially note the piping details on both the inlet and outlet nozzles of all involved PSVs – you will have to confirm the maximum 10% inlet loss on the PSVs as well as any discharge header back pressure. This is hard, expensive, and time consuming work; but it must be done since the existing P&IDs are essentially “worthless” and would take too much time to update and As-Built. Try to obtain a team effort on the field analysis by soliciting advice and comments from operators as well as instrumentation technicians. Do a mini-Hazop with the operators and instrument techs as well as with any process or production engineers as you can obtain. This mini-Hazop is done prior to identifying the worse case scenario and will help you identify the same in a quicker time span since it will be developed out of a process “summit meeting” involving the real experts in your process – the operations and instrumentation team. Presumably you will do all this work utilizing all the conventional MOC (Management of Change) guidelines set out by organizations such as OSHA in the USA. This will force out the real truth (and shortcomings) of the existing situation in a documented manner and it will help you out politically and professionally in the long run. This the legally and morally correct manner of rectifying such defective and potentially hazardous situations in existing plants and will empower you to take the necessary steps and decisions to make your workplace safer for your workers as well as for yourself.

If you are unable to carry out the above work because you lack the resources or experience, then I would repeat what I originally stated: Obtain the services of an experienced and capable process engineer who will deal with the problem in a correct and documented manner (as per MOC).

Obtaining an actual analysis of a sample case is, in my opinion, next to impossible because operating companies do not want to share this type of proprietary information due to process secrecy guidelines. You are correct in that API 520 is directed to “professional process engineers who can analyze the different scenarios”. But isn’t that what is required in order to ensure that the process is 100% safe and protected against harm to its operators and equipment? As I previously stated, this is a very serious safety situation and it requires a very serious application of experienced talent and efforts.

Like most of the Forum members, I wish I could help you more by being there and directly helping you apply the solution to this very important problem. However, that can’t be done and all we are left with is general recommendations through the Forum.

I wish you the best of good luck in what confronts you and that you obtain the safe environment you and your workers deserve.


Art Montemayor
Spring, TX

 

[farzinchemical]

thankyou very much montemayor,

i realize that you have spent a lot of time answering my rather preliminary questions and i want you to know i am very grateful in this regard.

i hope that i would someday have the privilede of being a part of a team which is conducted by an experienced engineer like you.

one last question. can you tell me more about MOC. what does it talk about? does it provide a systematic way of attacking technological problems? where can I start learning about it?

thanks again.

 

[Montemayor]

farzinchemical:

Being of some help or guidance in this type of common, but important, problem is thanks enough for me. I know that other members on this Forum feel the same.

Management Of Change is probably one of the most important – if not THE most important function of an operating plant engineering group. As you have noted, it is an organized and systematic method of attacking AND DOCUMENTING day-to-day changes made in the operating procedures and methods of an operating plant. By documentation, the MOC method ensures that the safety and proper operation of a plant is never jeopardized without knowledge or advisement to all interested and involved personnel. Many process plants have developed their own MOC forms which spell out and detail the indicated steps that a plant engineer must take and complete in order to make or impose a change in the plant's procedures. Any change made must be recorded in the Instrument Of Record (which is the P&ID of a process plant) and/or its written operating instructions –complete with face-to-face training of all operators and maintenance personnel. One of the main attributes of an MOC is that it makes the plant management team the virtual pro-active leader in accountability regarding changes made because they ultimately have to sign off on the approval process before a change is incorporated.

Any change to an “as-built” P&ID constitutes a process change – as does any modification in operating procedure or safety guidelines. There are standard identifications as to what constitutes a “change”. Certainly, your PSV project will result in a change if the resulting correct PSV is not an exact “change in kind”. All of this process is designed to build in safety in all plant operations and is a positive, legal, and moral backup for all plant personnel – including management, who are ultimately responsible for all that happens in a plant – be the results good or bad.

This is an important question that you pose and it deserves a lot of more, detailed attention and comments other than what I have offered. I am hoping that the rest of the Forum members can add they valued experience and help in this important matter that affects all of us, sooner or later.

I hope these comments are of help to you.


Art Montemayor
Spring, TX

 

[Guidoo]

farzinchemical:

First of all I want to fully support the arguments made by Art Montemayor. His contributions are (as always) very valuable.

For as far it is possible to check, your calculations for blocked outlet case seem to be OK.

My question is, why didn't you consider the fire case? Normally, fire case will be considered for a vessel containing liquid hydrocarbons...

 

[farzinchemical]

thanks montemayor and guidoo,

actually i have studied the fire case as well. assuming:

1. oil outlet blocked.
2. oil inlet blocked.
3. gas outlet blocked.
4. liquid level at normal operating level.
5. conservative latent heat of vaporization, 50 btu/lb(hysys calculates 300 btu/lb for the first 5% vaporized).
6. each of the two PSV's should be able to relieve the load at relieving conditions.
7. backpressure 8.5 psig.
8. relieving pressure 450 psi plus 21%.
9. critical flow conditions prevail.

but the load didnt differ with the blocked outlet case greatly.

the vessels dedicated for the job have been made in america and have a 6" outlet. how do the manufacturers know what size would be required. does this mean that the PSV should most probably have a 6" inlet flange (based on their preliminary calculations)?

 

[Montemayor]

farzinchemical:

Vessel fabricators - unless they are also the main engineering design and construction (E&C) contractor that designed your facility - don't have any input or responsibility for selecting the size or location of the PSV nozzle on pressure vessels. These nozzles are sized and specified by the E&C process engineer(s). The fabricator merely follows instructions or specifications when fabricating the ultimate vessel. If the process is designed correctly, one would expect that the existing PSV nozzles are the correct, sized, and specified nozzles and should suffice. But without the hard documentation, one can't be sure. And that is the existing situation defined previously. When you lose or misplace the original plant design calculations and documentation you've lost a large chunk of your expensive investment. Now, to continue to operate safely you are forced to essentially duplicate the effort once again.


Art Montemayor
Spring, TX

 

[TrevorP]

farzinchemical,
I have very little to add to Montemayor's excellent and insightful contributions.
The only one thing that I could add relating to the relief that you might not have considered until now is two-phase relief. If the vessel has an inflow of oil and gas, and all the outflow can be blocked, then it may be a requirement to relieve the oil and the gas at the same time. This can make a large difference to the relief device size. Some guidelines on how to calculate two phase relief is discussed in thread798-52877 and elsewhere.

Another place to look that might give some assistance on relief valve requirements is the plant HAZOP.

 

[jay165]

farzinchemical:

To expand on Art's point about complete information, one possible case you did not mention is gas blow-by from an upstream separator. Your vessel sounds like a LP or IP separator - do you have an HP separator upstream? If so, does the liquids from that separator go to the vessel in question? If the liquid dump valve gets stuck wide open or if the valve stem breaks, (and the liquid level in the HP separator empties)can the PSV on your vessel flow the volume of gas that passes across the dump valve?

I've found that this case gets overlooked - especially when there is an increase in production levels or changes in the GOR of the incoming stream.

You may also want to check out API 14 C to see if your oil and gas facility has all the recommended (and required in the US) safety devices.

 

[farzinchemical]

hello again,

i did as montemayor advised. we made another visit to the site and realized that the PSVs actually installed were not at the same size written on the non-asbuild P&IDs. they were much smaller. of course non of them were 0.44 in2.

I repeated my calculations with 300 btu/lb latent heat of vaporization. this was the heat of vaporization for the first 5% of the liquid vaporized. the orifice sized obtained were still small. but when i repeated the calculations with 50 btu/lb, which is recommended by API-520 when the latent heat is unknown the results agreed well with the sizes of PSV's on the existing installations.

my conclusion is that maybe the designers who designed the exsisting surface facilities did not have powerful simulators like hysys at that time and therefore used the 50 btu/lb assumption.

I know that consistency between the calculation results and what has been selected 30 years ago does not mean that my calculations and the selected worst case scenarios are correct. just wanted to inform people who are interested in the case.

i'm still investigating the case. i must add something more to jay165's remark. there is no upstream separator. this is a first stage separator.

trevor P, you might be right. i have never examined the possibility of liquid escaping from the PSVs. we have emergency shut-down valves which are being controlled by the level switches on the separators. should i really investigate this possibility?

thanks everyone.

 

[TrevorP]

Farzinchemical,
In answer to your question. As a rough approximation, if this is a hard-wired trip that is properly maintained and tested at 3 monthly intervals, I would expect a reliability of about 95%. That means that 5% of the time, when called upon, it will not work. If you have 2 independent switches and 2 independent valves, this would increase to about 99% (Maybe a bit more depending on the hardwire circuit). In either case, I would look at this possibility as additional to your current relief contingencies. It would certainly do no harm. Liquid on its own might not have an impact. Liquid and gas relief at the same time almost certainly would.

 

[hotdipper]

The following link is to a paper published by the Chemical Safety Board. It gives some excellent examples of what goes wrong when MOC is not implemented properly. The CSB web site is actually a good site for learning what not to do and the catstrophic results of poor implementation.

http://www.csb.gov/safety_publications/docs/moc082801.pdf

 

[Montemayor]

I don't know about others, but I had difficulty in trying to get to this .pdf document. What I wound up doing is going to

http://www.csb.gov/safety_publications/

and then clicking on Management of Change. This took me directly to the subject document.

I certainly agree with the CSB report on these 2 tragic incidents that could easily have been avoided. These example incidents clearly point out the need to implement MOC procedures in a process plant.