Desuperheating in PRO/II 1

[robot6]

What's the best way to model a desuperheater?

Difficulty arises if...

1) you try to control against "dew temperature," because you can add slightly more water and the temperature doesn't change -> solver doesn't solve

2) you send the outlet of the mixer into a phase separator and look for a small amount of water, again solver does not act robustly

3) it seems the best solution I can find is mixing the superheated steam and water first, and then looking for a vapor phase fraction of 0.99 (0.999 is not robust)... But this really has the same problem as #2, and this solution to me only "seems" more robust that #2, (I can rationalize a few reasons why it might be this way)

The inherent problems is, of course, the discontinuous derivative (phase change) that is part of this physical process.

The _motivation_ of solving this problem is that I need to compare various regimes of operation with the goal of maximizing system efficiency. If I use too much water, I'm carrying energy away by heating the water, but if I use too little, then of course, downstream processes don't act as efficiently.

Thanks for whatever insight you might have, be it in HYSYS or PRO/II or BRE or whatever.

 

[PaoloPemi]

there is low cost and free software which does that
for steam, you need a good source for water and steam properties (water and steam enthalpy plus saturation pressure) see for example IAPWS tables, I do that in Excel or Python with a thermodynamic library (Prode),
you need a solver or use the procedures in the software,

https://beta.spiraxsarco.com/learn-about-steam/desuperheating/basic-desuperheating-theory

 

[robot6]

Ah, yes, you can do it in excel and I have done it there a ton of times. In fact I actually coded the IAWPS equations into visual basic so that I could get away from tables.... which crash the solver in excel because of the discontinuity betwen points.

The point of the posted question (as well as why I used a solver in excel) is that this is just one tiny unit op in a pile of so many other operations. This is not actually desuperheating steam from a steam header, that's a one-off problem that takes less than 5 minutes. This is finding the optimum operating point for an overall process, which contains at least a hundred unit operations, 3 recirculation loops and ... somewhere in the process is a desuperheater that actually doesn't get steam off a steam header, but is prepared from a compressor... on a flow that varies from another unit op, and is controlled by demand from a downstream variable... So no, it's not really that easy to do in excel.

Thanks for your reply!

 

[PaoloPemi]

if you wish to create a model for a simulator consider the possibility to write some script/procedure with the tools included,
I prefer thermo libraries which are, in my opinion, more flexible if you need to define new models etc. the library which I use has a generic model for a desuperheater but it would be easy to create one with a script or compiled, native code,
and, yes, you can even solve large problems... but I admit that the graphical interfaces and wizards provided by most simulators can help

 

[Iomcube]

In most of simulators there is an adjust block. PaoloPemi rightly said why reinvent the wheel when you have thermo-libraries. In DWSIM or COCO (opensource simulators) for instance I will
1] Feed superheated steam
2] Water at say 20C & 3barg (must be greater than 1]
3] Mixer Unit operation for 1] 2]
4] Outlet product
5] Adjuster to iterate ....
Adjusted var: Water flowrate 2]
Set point var: VapFrac of 4] say 0.99

 

[georgeverghese]

Another way would be to
a) DEFINE another stream as the exit stream from the desuperheater
b) Run a DEW TEMP ( or DEW TEMP ) FLASH operation on this new stream
c) Use a controller to VARY water flow to the desuperheater to meet the SPEC that the TEMP reported by the dewflash DIVIDED by the TEMP of the original stream is 1.0 within a TOL limit you can set to suit.

Pro II has to the tools to get most of your work done if you dig around in the keyword manuals / casebook applications for clues.

 

[marcsant]

PaoloPemi is correct,
for a pure fluid you know inlet H, saturation conditions (final conditions) and latent hest,
you can solve quickly, it's simple with a thermo library (as Prode) or Steam Tables etc.
in your case the complexity arises from the fact that simulator doesn't expose the internal API (as thermo libraries or open source products as those mentioned by Iomcube do),
hence you need some trick... but for those tricks probably it would be better to chat with developers...