When we tried to simulate a compression process applying the Peng-Robinson-EoS to a gaseous mixture or even to a pure gas, we encountered severe difficulties with exit pressures of the compressor being set as high as 150 bar or higher.

Is anythin known about an obvious - or not so obvious - upper pressure limit when using real-gas equations of state?

Thanks for helping us. Regards,


Markus

It is ok in CAPE-OPEN to return an error condition if the answer cannot be found. However, if the simulation actually needs to peform calculations at these conditions it will of course cause the simulation to fail.

Presumably the problems you are facing are of numeric nature. EOS models generally have an asymptote towards a constant (excluded) volume at very high pressure. If your EOS fails to converge, you could possible check for this very high pressure and return the asymptotic values.

In this very high limit it is not likely that you still have a vapor phase though (if it is supercritical however labelling the phase as vapor is ok).

Pressure and temperature do not have (obvious or otherwise) upper limits. More-over, the compressor does not know it is dealing with an EOS model if it uses CAPE-OPEN thermodynamics.

We are trying to setup a simulation of a dehumidification process for natural gas refinement. This process involves two or three stages where a mixture of real gases undergoes an increase in pressure by a factor of two or more driven by a device called a turbo compressor.

As described before, we failed getting useful results, i.e. the calculations did not converge.

Therefore, we started to reduce the problem step by step, finally arriving at a single-step compression of a real gas consisting of only one component. The pure substance we have reduced the mixture to is methane (pc = 46 bar, Tc = 190 K).

Now, starting our compression at p1 = 65 bar and T1 = 297.85 K, we do obtain meaningful results, as long as the pressure at the outlet of the compressor is set to p2 = 116 bar (--> T2 353.78 K) or lower. The identical calculation, however, fails to converge, if the exit pressure is set to any value equal to or greater than 118 bar. The Peng-Robinson Equation of State was used.

First, we would be eager to understand why the calculation failed even if it is the pressure and not the specific volume that was prescribed at compressor exit. Secondly and perhaps more importantly, we need a way to get around this problem since the compression of natural gas from 65 bar to over 140 bar is only one step in a whole process that should be treated preferably using a real-gas equation rather than the ideal-gas law.

Is there a possibility to switch gas laws within the simulation of a complete process? Are results still trustworthy and would this kind of strategy to get around a problem be common?

Many thanks in advance.

Which simulation package are you using? Which software is providing the thermo (who calculates the PR equation)? Which package implements the compressor?

(I presume that some of the above are from different vendors, or CAPE-OPEN would not likely be involved).

As it seems, we are making progres in solving this problems. My student told me that the single-step compression now runs fine and that the difference to our previous attempts was that he did not choose all of teh properties by himself but rather let the program using the ones ther were set as default.

Thank you so much for your guidance!

Markus

I would also be interested in knowing which simulator and which CAPE-OPEN process modelling component is involved. With respect to natural gas, the most obvious CAPE-OPEN component I know is the GERG-2004 Property Package developed at Ruhr Universität. It has been used in Aspen Hysys and PRO/II with success on full scale process plants.