Safety Relief Valve Sizing

[silvz71]

I am trying to size a relief valve on a small mixing tank for a pilot plant operation. The worst case is for a very viscous material (65,000cP) and is blocked in fire case. Everywhere I have looked about finding the capacity of valve wants to know the orifice area. I thought that was what I was trying to determine! Can someone please help me determine the necessary GPM that the relief valve needs to accomodate? Any help at all would really be appreciated....thanks!

 

[dcasto]

Trail and error it, or write your own program to solve backwards.

Makesure the equations match the stated size in the US its either ASME or API, each one gets the same results but the eqautions/orifice sizes are diffferent.

 

[EGT01]

dbaylor,

In order to size a relief valve you need to know the "required relief rate" and conditions. You take that information and plug that into the relief valve sizing equations such as provided in API RP-520 "Sizing, Selection, and Installation of Pressure-Relieving Devices in Refineries, Part I" or as provided by a particular valve manufacturer. When it comes time to purchase the valve, complete a relief valve specification sheet such as found in API RP-520 and give that to a valve manufacturer/supplier and they generally will perform the calculations to verify your request before giving you a quote.

Now you say you know the worst case but I have some question that you do because if you have not been able to size the valve yet how do you know which overpressure case is the worst case. The exception, I suppose, would be if blocked-in fire is the only case.

You state that everywhere you look about finding the capacity of a valve you need to know the orifice area and that is a correct statement. The capacity of a valve is determined not only by the orifice area installed in the valve but also by other aspects of a particular valve design. But don't confuse the capacity of the valve with the "required relief rate". "You" have to determine the required relief rate and conditions before you can select a valve of sufficient capacity and that is compatible with your service.

Off hand, I'm not sure what "relief valve" would be suitable for relieving a liquid at 65000 cP. I don't have experience with such material but I suspect the viscosity correction factor that would apply for such a fluid would suggest that some other type relief device would be indicated, possibly a rupture disc instead of relief valve. The viscosity issue is something you probably need to investigate straight away with a relief valve manufacturer such as Farris, Consolidated, Anderson-Greenwood just to name a few.

Okay, so as far as getting help to determine the necessary "required relief rate" and relief conditions, that is something that cannot easily be provided through a forum. Your best bet would be to seek the assistance and direction from an experienced designer at your location. I could make a guess as to what you might need to do but, without knowing all the details of your facilities, most likely it would be an incomplete guess at best.

If blocked-in fire was the only case and, if your system is liquid full and, if the material that was contained in your mixer would not vaporize at the relief set point and, if the material was not prone to decomposition at elevated temperatures, then you might get by with determining the required relief rate as if for liquid expansion due to the heat input from the fire. Hopefully you get my point about all the "ifs".

If you want to start learning relief system design, my suggestion would be to start with reviewing API RP-520 (mentioned above) and reviewing API Standard 521 (formerly RP-521) "Pressure-relieving and Depressuring Systems". API 521 includes discussion about various overpressure cases that should be considered in order to determine your worst case event and does give guidance for determining heat input from fire exposure as well as a means to evaluate the liquid expansion case. The AIChE publication "Guidelines for Pressure Relief and Effluent Handling Systems" is another good reference.

 

[silvz71]

Thank you both for you posts. I guess I did use the wrong terminology for this problem. I am having trouble determining the "required relieving capacity" for the valve. I have no idea how to predict what this material will do in an overpressure or fire case. So, I don't know what the relieving capacity would need to be to keep the system protected. Like I said, I've seen lots of resources that determine the orifice area given required relieving capacity. I'm just wondering how, in these examples in manuals or books, they determined that the required relieving capacity is 100 GPM for example instead of 150 GPM. I apologize for spinning my wheels on this one, but I'm really having a hard time wrapping my head around it. This is a new application of my skills and I'm still learning! Thanks for the help!!

 

[pleckner]

I and others can go into a long description of what you need to know to be able to do this stuff but to answer you question more directly, you must have, as a minimum, an experience Process Engineer looking closely over your shoulder. In the absence of an experienced Process Engineer then your company should be hiring a firm that has people that specialize in these type of calculations.

Sizing the PSV is the easy part as you've found out; it only takes inputting values into typical equations. Finding out which equation to use and what these values are is the hard part. Depending on the system, you need to have a good working knowledge of pumps and how to read a pump curve; you need to know hydrauilcs, i.e. pressure drop and line sizing; you need to have a good working knowledge of thermodynamics, e.g. vapor-liquid equilibrium; you need to have a good working knowledge of heat exchangers, e.g. maximum Q based on the clean coefficient; is there a temperature pinch?; you need to have a good working knowledge of separations systems, e.g. distillation, stripping and absorbtion systems; and etc.

Having a novice doing PSV relief sizing is just asking for trouble as this is not something you can afford to "get wrong". I don't think I really need to tell you that if you do get it wrong, there is a substantial risk to not only material damage in your facility but loss of life as well.

You would be well advised to follow the recommendation given to you by @EGT01 above in the last paragraph of his post.

 

[Ashereng]

I have no idea how to predict what this material will do in an overpressure or fire case.

Yup. A process engineer (often a chemical engineer) will need to determine what happens to materials as they start to heat up in a fire case. Most "fluids" are also not a pure substance (eg oil) and does different things at different rates/temps/pressures/etc.

I would suggest going back to whoever asked you to do this and let them know you can't do it, and get some help to determine the relieve capacity required.

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[silvz71]

Well, sadly I AM a chemical engineer. However, I just recently graduated and, well, I'm pretty green. I've had good experience with lots of things, but just not this. (Plus the company won't tell us much about their material) I'll take your advice and see if I can work with someone else on this one. Thanks a bunch for everyone's help. This is a great forum!

 

[Ashereng]

dbaylor,

Not "sadly". It is only sad that your boss put you in this situation. At least you knew enough that you needed help. Hopefully, you will find it.

There are lots of things you will not know in your career. The trick is to recognise when you don't know something and need help. The fact that you can recognise this is saying quite a bit.

Good luck.

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[dcasto]

Start with the MSDS, it will tell you some of the properties you need. You maybe able to fill in the blanks too.

 

[silvz71]

Thanks for everyone's help and advice on this. I actually attended the API pressure relieving course last week, which was very helpful. I think I have the procedure/concepts down, however I still have a question I was hoping someone could help me with. The material I'm dealing with is very viscous (65,000 cP). So, when I use the iterative process to adjust for viscous flow, the value of Reynold's number quickly decreases. (R=(2800*G*Q)/(mu*sqrt(A)) So, each time I calculate an orifice size, then choose the next closest standard size valve, and recalculate. I end up getting an even larger area and this continues and the area increases to ridiculus areas. Does anyone have some advice on this? This valve is going on a 27 gallon tank in a pilot plant operation, so I can't imagine this valve being too big. I hope I wasn't too confusing....if anyone needs clarification, please let me know. Thanks!

 

[pleckner]

You might want to get a hold of the following article:

Darby,R. and Kamyar Molavi, "Viscosity Correction Factor for Safety Relief Valves", Process Safety Progress, Vol. 16, No. 2, Summer 1997.

This is a publication of the AIChE, Safety Division.

 

[EGT01]

dbaylor,

It may have become lost in my lengthy previous post, but what you're seeing is precisely one of the things I was trying to warn about.

Though I don't have the specific article PLeckner refers to, I have seen that article referenced in many other papers so certainly it wouldn't hurt to get a copy and I'm sure that will help to better design a relief valve for your system.

But, considering that you are only talking about a pilot plant with a relatively small volume, I wonder how important it is to have a reclosing type relief device and you might be better off with looking at a non-reclosing device such as a rupture disc.

For more information on rupture discs, one very good reference can be found here...

http://www.cheresources.com/asiseeit3.shtml

...you might recognize a name there.

By the way, I've often wondered about those API Pressure Relieving courses. I see that through Equity Engineering, there was a 3 day course June 19, is that the one you attended? What did you think about it?

 

[silvz71]

The three day course by Equity Engineering on June 19-21 was the class I attended. I thought it was very helpful and actually kind of enjoyable. The class was made to be interesting and the instructors were very easy to listen to for 3 days. I would recommend it to anyone thinking about taking it.

 

[Tweec]

What topics were covered at the API seminar? I would like to learn more about real-world applications of API 520.

For instance, I have precious little data for sizing a PSV on a knock-out drum. In the absence of hard information, how would one go about calculating the worst-case flow rate? The Process Engineer provided the S/N on the upstream PCV (All written data lost in flood; mfr could only provide the maximum flow rate for the valve series.)

Is it reasonable to use Bernoulli to find velocity based on the pressure differential of the upstream vessel's PSV set pressure and the vessel in question's PSV set pressure? Then Continuity to calculate Q, using the area of the PSV nozzle?

 

[silvz71]

The topics the covered, in powerpoint form were (they gave us printouts of the presentations):
Introduction, Basic Terminology, PRV Design and Operations, Codes and Standards, PRV Sizing and Selection, Two-phase flow sizing, Rupture Disks, Installation of Pressure Relief Devices, Relief System Design Philosophy, Overpressure Scenarios, Atmospheric Storage Tanks, Power Boilers, MOC and Documentation, Specifications/Data Sheets, PRV Inspection, Testing, and Repair.

Nearly each section had "sample problems" that we worked individually then discussed as a class. Also, the course description says that you should bring API 520 P1, API 520, and API STD 2000, however I only looked at these once during the course and I really didnt need to do that. They give you all the equations, tables, and figures they refer to in the discussion and problems. Of course you will eventually want these as a resource, but I had to hurry to get them because I registered so late, and they really were not necessary.


As far as you question goes, I can't definitively give a suggestion. It sounds like your assumptions are reasonable, however I'm still pretty new to this and I don't want to make a recommendation. Maybe someone with more experience can comment.

 

[silvz71]

If anyone has any other suggestions for articles, I would appreciate the reference. I still have access to some databases through Purdue, so being able to save articles like that now would be great. Thanks!