Controlling Pressure to Condense

[toshibe]

Hello, everyone.

I'm now trying to simulate a very simple liquefaction process of Ammonia as attached.
It's just single process but it is necessary to adjust manually the pressure so that the compressed
vapor is fully condensed in the cooler(condenser) according to the given condensing temperature.
Then, I have involved one set of "MeasureUnit" and "Controller" to obtain the condensing pressure
automatically. The connection of information streams and setting of the parameters looks good
but the iteration behaves as if nothing will converges. I have tried various parameters relating to
the iteration but nothing is improved. There seems to be still something wrong.

Actually I have tried the similar process of LPG and confirmed much better behavior of the
controller. It has been noted that the initial guess of the controlled variable is a key point to get
better convergence. Then I have tried the same manner for Ammonia but failed to get better
behavior.

Does anyone guide me to proper setting for the "Controller" with Ammonia?

Thank you in advance.

[jasper]

The controller is doing what it is asked. The vapor fraction is at 1.

What you are looking for is the dew point pressure at 40 C. You could measure the dew point pressure and actual pressure of the stream (or dew point temperature and actual temperature), subtract the two (information calculator) and control the pressure at the compressor so that the difference is zero. As the HeaterCooler does not actually perform a vapor fraction flash, the result will not be exactly at the phase boundary.

[toshibe]

Thank you very much for your reply.

Accordng to your guidance I have modified the flow sheet and confirmed the behavior what I want.
The modified flow sheet is attached just for reference.

Again thank you for your kind assistance!!

[jasper]

Glad to hear it.

[jasper]

(you used the bubble point pressure - is that intended?)

And even simpler solution is to remove the information calculator and controller, and immediately connect the dew point (bubble point?) pressure measurement as input parameter. I attached it for your reference.

[toshibe]

Thank you for your additional guidance.

Yes, the bubble point has been intended as the preparation for the next simulation with multi component system such as LPG.
My concern is the total refrigerating capacity (thermal load) shown by the energy stream 11 and it is necessary to get the full condensation in the condenser (HeaterCooler_4). So, the bubble point is the better index in the case of multi component system.

In this regard your simpler way with direct connection of the dew point (or bubble point) does not get the full condensation in the condenser. Please have a look at the vapor fraction in the stream 5 or 6.
Then I have implemented an information calculator to give the pressure just slightly higher than the dew point (or bubble point) as shown on the attached file.
Now I can get the total refrigerating capacity with the simpler way you guided.

Thank you again.

[jasper]

I had pointed that out already:

As the HeaterCooler does not actually perform a vapor fraction flash, the result will not be exactly at the phase boundary.

You cannot get a stream to be exactly at the phase boundary with a TP flash. The TP flash in this scenario will put you as close to the phase boundary as the convergence tolerance of the flowsheet (because there is a recycle involved - without the recycle it would be as close at the convergence tolerance of the thermo - or at best machine precision on T and P.

The only way to be exactly at the phase boundary is to use a flash calculation for which the specification is that it should be exactly at the phase boundary, so a specification with vapor fraction = 1 or 0. As the other specification you have is temperature, this would be a T-VF flash - not many units that perform one of those. You could instead use a flash unit and specify vapor fraction and pressure, where pressure is the dew point pressure at given temperature (as determined now). This would give you a stream close in temperature to your target (within flowsheet tolerance) and exactly at the phase boundary.

[toshibe]

I would convey my sincere thanks and appreciation for your informative advice.
I have well noticed my lack of depth of understanding through this issue and then I will try again with your advice.
It is very nice opportunity for me to discuss with you.

Many thanks for your time and help.

[jasper]

Glad to be of help.