[Bug] COSTALD liquid density model results are not consistent with benchmarks

[inemov]

Describe the bug
When evaluating hydrocarbons+nitrogen mix density using COSTALD model on top of PR EOS, the results are not consistent with benchmarks:

• DWSIM PR EOS + RACKETT
• Thermopack (https://github.com/thermotools/thermopack) + Chemicals (https://github.com/CalebBell/chemicals)
  Surprisingly, when DWSIM PR EOS is used with EOS liquid density model, the results are close to the above benchmarks, but different from results obtained using Thermopack PR EOS. I.e. DWSIM PR EOS + RACKET produce density very close to pure DWSIM PR EOS.

My guess is that when EOS liquid density model is selected in DWSIM, it is actually using COSTALD model. And when COSTALD model is selected it is doing something completely different.

EOS Test.zip

To Reproduce
See attached zip file.

Expected behavior
Expect COSTALD results closer to RACKETT and EOS density model results different from RACKETT.

Screenshots
Please note the trend names are in "BACKEND"-"FLASH MODEL"-"ENTHALPY MODEL"-LIQUID DENSITY MODEL" format:

Desktop (please complete the following information):

• OS: Windows 10
• Version DWSIM 9.0.5