8 replies to this topic

[Soothsayer]

What do you recommend for sizing control valves, true or pseudo critical pressure? What do you think for hydrocarbons (diesel, fuel oil, HVGO...)?
The difference between both values may be significant in some cases, which affects the control sizing calculations.

Thank you in advance for your response.

[ChemEng01]

I normally use the True critical pressure as it is the more realistic value for critical pressure calculated thermodynamically by HYSYS.

 

The Pseudo critical pressure is just calculated using linear models. The only time you need this is if a vendor asks specifically for the pseudo critical properties.

 

For valve sizing definitely use true

 

cheers

[marchem]

which kind of condition and procedure ?
if it's for determining choked flow condition
with ISA 75.01 do not forget it is applicable to pure fluids,
for mixtures you can get large errors,
there are papers (Samson) discussing HNE method,
another possibility is HEM,

it is not difficult to get the TRUE critical pressure
and the TRUE vapor pressure directly in EXCEL
with a library such as PRODE PROPERTIES,

but I have compared results for liquid mixtures
with true (critical and vapor pressure) and
average values (i.e. pseudocritical and
average vapor pressure) with HEM and HNE
and there are quite large differences

perhaps it may depend from the fact that parameters
as liquid pressure recovery factor etc.
are not easily available and have a large impact on results,
in these cases I prefer to ask the manufacturer
for the proper selection.

Edited by marchem, 14 November 2013 - 04:51 PM.

[Sathya R]

Hi,

In Hysys.

The pseudo critical properties are simply the summation of the "apparent" contributions of each pure component. For example, the pseudo Tc of a binary mixture (component 1 and 2) would be calculated as Tc=x(1)*Tc(1)+x(2)*Tc(2).

The true critical properties are the real critical properties of the stream determined thermodynamically by satisfying both the quadratic and cubic forms in the expansion of the Helmholtz free energy as a function of the mole numbers as zero at a critical point. Details of critical point calculations can be found in the paper by Heidemann and Khalil (AIChE Journal, Vol. 26, No.5, p769-779, 1980).

On the other hand the values given in by phase envelope is more realistic as it is constructed by a marching algorithm that constructs one branch of the envelope (e.g., dew line) starting from a "low pressure" point, e.g., 1 atm.
As it generates more points (at higher T and P), it checks to see how close it is to the critical point (a quantity related to summation of ln(Ki), where Ki is the K-value of component i). When it thinks it gets close enough to this point, it switches over to the other branch (i.e., bubble line).

The actual critical point lies between the switch-over points. The algorithm does not try to find the actual critical point because it is a singular point as far as the algorithm is concerned.

On the other hand, the critical properties utility solves the condition of criticality (second derivative of Gibbs free energy equal to zero). As this is computationally expensive, it is not used in the normal flash calculation or in the phase envelope utility.

 

[ChemEng01]

If you were to use property table in HYSYS for  HEM method for control valve sizing the mass density and mass enthalpy with reducing pressure as calculated by HYSYS to determine the mass flux would use the true critical properties anyway.

 

The true critical properties are thermodynamically modelled by HYSYS and can differ slightly depending on what fluid package you use.

 

Easiest way is to convert your valve CV to an equivalent orifice size and use HEM.

[marchem]

JRudd,
note we are speaking of choked flow,
i.e. flashing fluid, not critical flow
i.e. speed of sound limit,
by chance do you mean a simplified
HEM variant used for PSV design/rating
as those based on Omega method
and equivalent simplified models ?
Anyway I do not know a HEM based on
true critical properties for control valves,
could you provide an example of a
control valve sizing with two phase flash
(not critical flow)
solved with that HEM methodology ?
I have a different software (PRODE PROPERTIES)
and I size / rate control valves in EXCEL
with the flash operations included in PRODE
(by the way HEM, HNE, NHNE are included in library),
but, as said in previous post, I found quite
large differences with different methods
compared with ISA adopting both TRUE and
average pc and vapor pressure,
this may depend from the difficult
to define the correct parameters
to model the valve.
Now, if you have a different method,
I can compare and report the results.

Edited by marchem, 15 November 2013 - 03:24 AM.

[ChemEng01]

I use the HEM method to calculate the breakthrough rate through a control valve for RV sizing. If the HEM method is used for RV sizing you have to use it to find the governing relief case from the control valve. Reason being is that equation such as the masonelian eqnsare conservative when calculating the CV so can lead to undersized RVs if the HEM method is used for RV sizing.

 

Control valve sizing using HEM:

 

1. Create property table in HYSYS for mass enthalpy and density at reducing pressure from U/S to D/S pressure. (To do this use expander at 100% efficiency to expand your stream at the u/s pressure and create property table from this expanded stream)

2. Plot mass flux vs pressure curve

3. Determine required orifice using max mass flux and appropriate discharge coeff

4. Use fig 14.3 from MILLER internal flow systems to determine a K factor (see att)

5. Use eqn 3-16 from CRANE flow of fluids to determine CV (see att)

 

Note that there are online calculators which will do steps 4 and 5.

 

Attached Files

•  Att miller and crane.docx   75.45KB   83 downloads

Edited by JRudd, 17 November 2013 - 04:19 PM.

[marchem]

if I interpret correctly your description
(points 1,2)
you use a procedure similar
to that used for a PSV
as I supposed,
you do not mention
which kind of flash operation is solved
in 1, 2 to create the table of values
and I can only speculate based on the
features of your software,
it would seem that you solve
for Hin=Hout at different pressures,
then calculate mixed volumes and
see if they are compatible
with valve CV (or port)
Note that I wasn't able
to access the attachments which
perhaps add some information
on your methodology

However, if you solve
for Hout=Hin at different pressures
you simplify the model,
for example you do not consider the
contribute of different
fluid velocities at inlet
and otlet which can be
important when densities
(as it is case with flashing)
are very different,
really here you should use
something like the
contant energy flash operation
available in PRODE PROPERTIES

another point to evaluate
could be the kind of model,
for example if you compare
HEM againt HNE in some cases
the second will show
very different values of
liquid vaporized, vapor
fractions, volumes, velocities
etc.
note that a constant
energy flash would
keep in account these differences.

but the main problem,
discussed in my previous post,
is that at design stage
we have little information
about valve internals
(i.e. single port, double port,
internal geometry)
and a generic model for Kv can fail
to predict correct values
(i.e. dp vs. fluid volumes)

I am not sure about which methods
the manufacturers do use for selection
in these cases,
I think the availability of relatively
low cost software as PRODE should
offer all the required features,
another possibility is to use a
simulator but of course that seems
more specific for plant's simulation.

Edited by marchem, 16 November 2013 - 05:28 AM.

[Hamedcheng]

Hi

 

How can indicate bitumen or vacuum bottom  Level  in bitumen tower (blowing tower)?

 

( bitumen in blowing tower is hot)

 

Thank you