[Bug] PVF / dew point calculation error

[andr1976]

Describe the bug Dewpoint calculation fails for the PVF flash for PR, SRK, GERG etc

Expected behavior Identify dew point temperature of -11 / -12 C'ish at 40 barg

Desktop (please complete the following information):

• OS: [e.g. Windows]
• Version [e.g. v8.0.0]

Additional context See attached file dewpoint_error.zip

[DanWBR]

@andr1976 does it fails due to an error or wrong temperature value? just to be sure

[DanWBR]

@andr1976 initial estimates are close to the trivial solution. the algorithm conditions are satisfied in the end, even with the wrong temperature. do you have any suggestion?

[andr1976]

The calculation runs but returns a temperature which is not the dew point at the specified pressure. Far from. Things start to go sideways above 30'something bar.

[andr1976]

[andr1976]

Ok one is with SRK but the same applies with PR. There is apparently a regression in the stability of the PVF/TVF algorithm

[DanWBR]

@andr1976 did you check this with latest code?

[andr1976]

Sorry, no haven’t had time yet

tir. 28. jun. 2022 kl. 20.34 skrev Daniel Wagner @.***>:

@andr1976 https://github.com/andr1976 did you check this with latest code?

— Reply to this email directly, view it on GitHub https://github.com/DanWBR/dwsim/issues/333#issuecomment-1169085992, or unsubscribe https://github.com/notifications/unsubscribe-auth/AN6EIDZ67JGT52PMIOYO3I3VRNAVDANCNFSM5ZG23CAQ . You are receiving this because you were mentioned.Message ID: @.***>

[andr1976]

@DanWBR , yes dew point is fine now, but the bubl-point is messed up now above 44-46 bar depending on EOS. I also think the critical point is way off. Above this pressure bubl point temperature suddenly jumps from -76 C to -10 C

[andr1976]

This is the envelope expected

[DanWBR]

@andr1976 the critical point is the reference to stop calculating dew and bubble points... it is very far from the correct one in this specific case

[andr1976]

Yes, I see. Thing is, this is natural gas (although a little to the heavy side), so this is something that could potentially be encountered quite often

[DanWBR]

This looks better (PR). SRK still fails to calculate the correct critical point. GERG will need a new algorithm as I don't have a general procedure to calculate critical points - only PR and SRK for now.

[DanWBR]

the chart title is wrong - it was calculated with PR.

[andr1976]

Looks way better now - I think PP is somehow defaulting to the last PP available, for the title, not for the calcs

[andr1976]

Could be nice to have Ian Bell on-board to do the critical point calculation method :-)
https://www.nist.gov/publications/calculation-critical-points-helmholtz-energy-explicit-mixture-models

[DanWBR]

@andr1976 I have an implementation of this method for quite some time already, but I've let it aside as it is slow and was not working. I'm revisiting it to see if I can get it working.

[DanWBR]

SRK: