Vapor release

[Wizard74]

Hi,

I would like to ask for help on a vapor release calculation.

A known superheated liquid at temperature T and pressure P, flows through a 50%-opened gate valve to atmospheric pressure in a 6" pipe. The liquid then flashes, travels 1m through the 6" pipe and is then released out of the system through an opening in a 0.23m long 2" pump suction strainer. The pump was idle. The duration lasted "t" minutes and all other properties of the liquid is known.

How can the amount released be estimated? Appreciate you guys' help! Any quick estimate or detailed calculation suggestion is welcome!

Thanks
Wiz

 

[Latexman]

Wizard,

If the superheated liquid at T and P adiabatically flashes to atmospheric pressure, is it all vapor or two phase?

Good luck,
Latexman

 

[Montemayor]

Wizard74:

How do you define a "superheated liquid"? Can you tell us where, on the T-S (or Mollier) Diagram, this liquid phase is located?

More specifically, what is the fluid's identification, its pressure and temperature, and what is the process you are operating? Or is this a hypothetical or academic question?

I'll await your response.

 

[sailoday28]

As Montemayor requests what is pressure, temp and liquid.
For information to others:
Superheated liquid exists in a metastable state. ie at 1000psia, Tsat 544.75F, liquid water can exist with temp as high as 620F---But again very unstable.

 

[sshep]

Please also detail the statement "then released out of the system through an opening in a 0.23m long 2" pump suction strainer." It sounds like this 2" section will be the primary resistance and the problem reduces to how much flashing flow can pass this point. Combining the discharge point geometry with the upstream pressure and clarifications of the superheat (i.e. outlet vapor fraction of the material flashed at atm pressure), then this is a problem for which you can probably come up with a reasonable answer. -sshep

 

[Wizard74]

Folks,

Thanks for your inputs. Perhaps I used the wrong word for "superheated". This is a hypothetical question.

To Latexman: it is assumed to flash all into vapor

To Montemayor: atmospheric boiling point is 250 degC, T = 280 degC, P = 2 bara

To sailoday28: thanks for the side info

To sshep: you are right, how much flashing flow can pass through this point.

Hope to hear from you all soon.

Thanks again
Wiz

 

[Montemayor]

Wizard74:

Now I'm really going to get on your corruption of the English language. What do you mean by "I used the wrong word for 'superheated'"? Are you trying to say that your liquid fluid is not superheated? Are you trying to say that your liquid fluid is saturated? If not, just what is the thermodynamic state of your liquid fluid?

If you don't identify the Thermodynamic state of your "hypothetical" liquid fluid, you can't be certain that it will flash. Hell, it could be supercooled liquid for all I know.

 

[Wizard74]

Montemayor,

The atmospheric boiling point is 250 degC, but the liquid is at 280 degC at 2 bara. This is the condition of the liquid and yes, I am trying to say that it is at saturated state. Sorry for the confusion.

Thanks
Wiz

 

[Latexman]

Wizard,

In your post you told me it flashes to all vapor. My gut tells me it probably won't flash to all vapor given the info. you gave to Montemayor. You also indicated to sshep it is "flashing flow", which is two phase flow. I think we are all a little confused now and the only way to save this post is to lay all your data out on the table, so to speak.

Good luck,
Latexman

 

[Wizard74]

Latexman,

Thanks for the feedback. What additional information will be required to calculate the flow through the opening?

The conditions supplied so far:
Atmospheric boiling point = 250 degC
Temp of liquid before 50% opened gate valve= 280 degC
Pressure of liquid before gate valve= 2 bara
Pressure of system/pipe after gate valve = atmospheric
Pipe diameter before and after gate valve = 6"
Length travelled in 6" pipe before strainer = 1 m
Pipe Diameter of strainer = 2"
Length travelled in strainer = 0.23 m

Total flashing to vapor is assumed (correct assumption?).

Thanks again for your patience.
Wiz

 

[mbeychok]

Wizard 74:

What is the name and/or chemical composition of your liquid ? Do you have the enthalpy of the liquid at 2 bara and 280 deg C as well as the enthalpy of the saturated liquid and saturated vapor at the atmospheric boiling point of 250 deg C ? With that data, you can calculate the adiabatic flash, if any, by using the following equation:

X = 100 ( H_s^L - H_a^L ) / ( H_a^V - H_a^L )

where:
X = weight percent vaporized
H_s^L = source liquid enthalpy at source temperature and pressure
H_a^L = flashed vapor enthalpy at atmospheric boiling point and pressure
H_a^L = residual liquid enthalpy at atmospheric boiling point and pressure

If that data is not available, then you can use the following equation:

X = 100 [ c_p ( T_s - T_b ) ] / H

where:
X = weight percent vaporized
c_p = source liquid specific heat
T_s = source liquid temperature
T_b = source liquid atmospheric boiling point
H = source liquid heat of vaporization at atmospheric boiling point

Until you have the above required data, you cannot say whether or not the liquid will flash or to what extent it will flash.

Milton Beychok
(Contact me at www.air-dispersion.com)

 

[Montemayor]

Wiz:

I, like the other responders, have given this post importance because it comes from a 2+ yrs forum member who is also a ChemE. You've had responses from some of the brightest and sharpest Thermo minds on these fora: Milt Beychok, Sailoday, Latexman, & SShep. I think you should level with them because you stand to reap a lot of valuable Thermo knowledge from them.

Don't take our responses too personal. We tend to be more serious with ChEs when it comes to Thermo because we expect more from them. We'll wait for your response.

 

[Wizard74]

Mbeychok & Montemayor,

Thanks a lot for the pointers. I guess I had not been a good student during my college days and been away from Thermo for years to remember how to tackle problems of this nature. Nonetheless, I am grateful for your responses and continued patience.

This is what I have done. I have initially assumed no flashing and then used the following equation to calculate the flow rate through the 2" opening:

Q = (pi * dP * D**4)/(128 * mu * L)
where
dP = pressure drop across the gate valve
D = 2 inch
mu = visocisty of liquid
L = length of 2 inch strainer pipe

This is Hagen-Poiseuille equation and assumes laminar flow and no flashing. However, I am aware that the assumptions are grossly invalid but this is about the fastest equation I know of.

Anyway, I have calculated the % flashed as 23% and 22% using the 2 equations from Mbeychok. I have read up Perry's and it seems that multi-phase flow calculations are in general complex. I have no experience in such calculations and would appreciate if you can guide me through this.

Thanks again
Wiz

 

[Latexman]

Wizard,

Do you have a reason for protecting the identity of the chemical? If you got the data from Perry's, it's public knowledge anyway. It would have been great if you had given us the data that led you to the 22-23% flash and the incorrect flow (probably) you calculated.

Please describe the 2" strainer pipe. Open end or closed end? Does it have holes in it? What size? How many? Etc.

Good luck,
Latexman

 

[Latexman]

Wizard,

My last post was interrupted, so consider this a continuation.

Yes, two phase flow calculations (liquid and vapor, in your case) are complex. The first step is to understand whether you have horizontal flow, upflow, downflow, or inclined flow. You may have some parts that are horizontal flow and some that are another type, like downflow for example. It would simplify matters *a lot* if your system could be represented by the predominant flow direction based on some criteria, like length of flow direction.

Despite the complexity and enormous research that has gone into two phase flow over the last 100 years, there is still substantial uncertainty in most of the correlations used today, so don’t feel that you are losing too much by characterizing your entire system by the predominant flow direction.

If you want the technical challenge, the next step is to determine the flow regime (wave, annular, spray, bubble, stratified, slug, plug) using the regime map for the flow direction you have. This will be trial and error since you have to assume a flow rate, calculate pressure drop, and see if it matches your problem. My 6th Edition of Perry’s has correlations for cocurrent horizontal flow (Lockhart and Martinelli) and vertical downflow for annular flow. For other flow directions and regimes, they point you to a reference document. This is not very convenient, huh? Again, it’s a complex matter.

It has become popular lately in the relief design area to use the conservative “homogeneous equilibrium model” method for two phase flow no matter what the flow direction or regime is. At least, that is my perception. My Perry’s doesn’t have this method in it, but I believe the 7th Edition does. If you google “homogeneous equilibrium model”, I got 214 hits.

You will have to decide how you will approach your problem. How accurate do you need to be? Do you just need to be conservative? Are you estimating a spill and release, and do you need to consider the amount evaporated after the spill occurs? These are all questions you need to consider. It’d help us help you if you could fill in some of the gaps. Like, what’s the chemical, why do you need the flow rate, what’s the flow direction, what’s the vapor density, liquid density, vapor viscosity, liquid viscosity, surface tension, change in elevation, number of velocity heads of the fittings, the flow area of the strainer, is the strainer submerged, if so, by how much, etc., etc., and etc. After you decide your approach, you can post your data and results from your calcs and someone will probably check it for you, or at least give you a ballpark gut check.





Good luck,
Latexman

 

[sshep]

Hey Wizard,

I think that you are on the right track with your conclusion of partial vaporization. Mbeychok has reminded you of the equations applicable- his second equation is typically the easiest approximation since Cp and H are usually known. It is truly an approximation since the liquid mass is being reduced during the flash and the generated vapor has a much different Cp- this should cause you to underestimate the vfrac. This may be argued to your advantage as conservative since you will then be overestimating the amount leaked.

My own flash for N-Tetradecane (chosen because BP=253C) from 2bar and 283C to 1bar gave 39% vaporization. Your 22% is possible given the right Cp/H. I am fortunate to have a process simulator readily available. This tool also gives me the capability to easily estimate an amount released, but it appears that you are working by hand. For a hand calc case I suggest a simple approximation might be to calculate the amount of vapor which could pass through a short 2" channel at choked flow. To get the total flow simply scale up from the amount of vaporization which is calculated from your flash. For example, if X is the choked vapor flow then X/vfrac is the total flow.

I realize that some purists might object to my suggesting such a shortcut, but it appears that you are doing some release estimates. The data available for release calcs is often so rough that the expected accuracy is far less than many other engineering calcs. My experience in such cases is that the most important requirement is to identify a defensible and reasonable basis in a timely fashion. Others in your organization should have some responsibility to understand and review your calculations. Your estimate is complete when a consensus of reasonability among the qualified staff is reached.

Just my thoughts- sshep

 

[Wizard74]

Latexman & sshep & those who had responded:

Thanks a lot for your help. I am estimating a hypothetical vapor release for a PHA study so all your pointers have been top-notch for me. I guess my approach will be conservative for the study. And since I'm new to the organization, I cannot be sure if it is appropriate to share more details than I already had so I hope you guys can understand.

Montemayor, you are right, I have learned a lot.

Take care
Wiz