Model Parameters for Data Prediction

Parameters delivered by DDBST

In order to provide reliable model parameters for the calculation of pure component and mixture properties, DDBST is for some time engaged in regression of the different data from the DDB. Currently numerous parameter sets are available. These include:

  • Equation of state (SRK, PR) binary parameters for the quadratic mixing rule (kij)
  • Twu alpha function parameters for the description of pure component vapor pressures via VTPR
  • temperature dependent interaction parameters for the gE models Wilson, NRTL and UNIQUAC
  • ...

The following figures show an example for the typical results of a simultaneous regression of different mixture data for the system cyclohexane - 1-propanol using the Wilson model:


Parameters delivered by the UNIFAC Consortium

The parameters for the UNIFAC, mod. UNIFAC and PSRK methods delivered by DDBST conform to the latest published version of the models. Since 1995 the methods are continuously improved and extended within a consortium. This work is currently sponsored by approx. 50 companies worldwide and has led to a huge improvement of all methods ( The public version of mod. UNIFAC contains 53 main groups and parameters for 650 group pairs. The consortium version improves more than 250 of these pairs and adds more than 650 further parameter pairs as well as more than 30 main groups of technical importance and is only available to consortium members.


Scientific Papers

2009    Recommended NRTL Model Parameters by Simultaneous Correlation of VLE, Infinite Dilution Activity Coefficients and Excess Enthalpy Data    Tochigi K., Rarey J., Gmehling J.    Journal    J.Chem.Eng.Japan, 42, 6, 376 380 (2009)