Relief Valve On Desalter - Fire Case 2

[prochem]

Hi all,

I am calculating a relief load of a liquid full Desalter for a fire case. I have a specific concern about flashing liquid at the time of relief.My question is can I size the relief valve based on Vapor relief after considering initial liquid relif due to volumetric expansion of water?What about the latent heat? The latent heat of the liquid(Hydrocarbon+Water) will change as the water and light ends would start vaporizing.Does any body have experience in this matter?

I will appreciate your suggestions.

Regards,

Prochem

 

[CJKruger]

prochem, there are two schools of thought.

Method 1. Estimate the heat of vaporization and vapor composition at say 20-30 wt% vaporization of the desalter contents. Calc the load from the wetted surface area. Size the relief valve based on this load. For header calculations, simply back calc how much liquid can pass through the valve, and report it.

Method 2. Size the relief valve based on the simultaneous relief of vapor-liquid mixture.

I would use method 1, but suspect the industry is slowly moving to method 2.

 

[prochem]

CJKruger,

Thanks for your reply.

I also believe that method 1 is more sansible approach as there will be some vapor space above the liquid (saturated).will be generated after initial liquid relief due to volumetric expansion.

However one of my co-enginners believe that the PSV need to be sized for two phase relief (approach 2) and use the HEM as mentioned in the API.

Regards,
Prochem

 

[CJKruger]

prochem, we recently finished a project which included a desalter psv, and used method 1. That being said, from previous discussion on this board (maybe Latexman or Pleckner), I got the impression that method 2 should be considered. If I do another desalter psv, I would read up carefully on the issue in the latest API 520/521 to ensure method 1 is still acceptable.

 

[pleckner]

The accepted practice for the fire relief of a pressure vessel full with a non-foamy, non-viscous and non-reactive liquid is to treat the relief as a vapor. If the liquid is foamy and/or viscous and/or reactive, then you must use the DIERS methodology to determine if you have two-phase relief or not.

One reference you can site would be: "Guidelines for Pressure Relief and Effluent Handling Systems", CCPS of AIChE, 1998.

Note that this "way out" from using two-phase flow relief becomes more acceptable as the relief device set pressure increases and conversely, decreases in acceptability as the relief device set pressure decreases. For atmospheric tanks or low pressure tanks, one must look very careful at not considering two-phase relief.

As an additional note, even if you consider only vapor relief in liquid full vessels, you should still consider two-phase relief for the sizing of downstream effluent handling systems, i.e. the relieving device can be sized for all vapor relief but not the downstream piping, catch tank, etc.

 

[prochem]

pleckner,

Thanks for your comments. I red through those references already.Finally we have decided to use HEM method for two phase relief.

This is a critical relief. Calculating mass flux is not an issue but initial calc shows very high instanteneous relief rate.Which doesnt seem realistic.

Regards.
prochem

 

[pleckner]

Is this a case of a small, volume to surface area, ratio? This could explain why you perceive the instantaneous relief rate to be high.

Also, it doesn't seem I addressed your question on latent heat. There are several schools of thought as to how to determine the correct latent heat for a mixture. Please go to the following link for a discussion I had on this subject:

http://www.cheresources.com/invision/index.php?showtopic=3742&hl=latent+heat

You can also do a search on the forum for "latent heat" for more information.

 

[lizking]

API 521, Fifth Ed., paragraphs 5.15.3.3 & .4 give some guidance on this issue. I believe alot depends on where the RV is located (i.e. top of vessel or lower in the liquid level). If it is at a high point, then the scenario is hydraulic expansion--2 phase--all vapor and the "the assumption of vapor-only venting is appropriate for relief device sizing." If the PRD is lower down on the vessel then "the PRD should be able to pass a volume of fluid eqivalent to the volume of vapor generated by the fire."

 

[salar1363]

Hi all,

Having similar projects, I'm totally confused with the required capacity of pressure relief valves on the top of Desalters, on previous projects, I have considered the required capacity to be equal to the inlet to the desalter regardless of thinking about fire case. Can you please tell me if this method is acceptable or not?? what is the purpose of having RV on the top of desaters ??
I have asked similar questions on this forum and would be so grateful if you help me on this issue.
Thank you

Salar

 

[pleckner]

@salar1363:

You can't just decide a particular relief rate is going to control the size of the PSV without going through your scenarios (as I and some others have brought up in your other post in this Forum) and calculating the relieving rate for each. And besides, if your place of origin falls within OSHA regulations (U.S.) or you have similar regulations, then you would be violating law unless you performed the complete analysis.

 

[prochem]

salar 1363,

pleckner is absolutely right.Check your local regs and find out.If you fall within OSHA, its about time you reconsider your approach. Its not only about violation of the law, we as a designer/engineer have moral responsibility to assure the safety and integrety of the facility too.

The possible overpressure scenario for your Desalter could be - However it totally depends on the specific system in question -(1) Blocked outlet / inadvertantly closure of any valve downstream or may be failure of an auto control valve.
(2) External fire
(3) Liquid overfill

These are few possible scenario considering typical Crude Unit Preheat train just to begin your thought process.Again you and only you are the best judge of your overall system.

Hope this help.

Regards,
prochem

 

[salar1363]

Dear Prochem,

Thank you so much for your kind advice. I tried to review ASME sec 8, and I will review API 520 and API 521.
Actually I'm not that much experience to design a whole package alone, and I always receive advice of my supervisor on these critical issues. However, I would like to know other opinions to be more educated.
a far as I could get, size of the valve(req. capacity) for fire cases would be smaller than Block flow scenario. Besides, for all pressure vessels, like desalters, having PRV for block flow is highly recommended. therefore, considering block-flow case will cover fire cases also ( set pressure does also). Now, it's my pleasure to have your technical comment, or please correct me if I'm mistaking.
I guess i couldn't make myself understood previous times. I've never considered rough value for capacity of PRVs. I was wondered if we needed any other consideration for fire cases in addition of block flow scenario.
Thank you again Prochem

Best regards

Salar

 

[prochem]

salar1363,

You can not assume that your 100% liquid relief due to blocked flow will be higher than your vapour / two phase relief due to external fire some other scenario.Hence do not take it for granted.

Your instantaneous vapour / two phase relief may be higher than the sustained liquid relief.

You have to establish / confirm by calculating relief reauirement for each individual cases.

Regards,
prochem