VLE Thermodynamic Consistency: Van Ness Test background
Test 2: Van Ness Test. This test23 is regarded as a modeling capability test.24 The test shows how a mathematical activity coefficient model can accurately reproduce the experimental data. In TDE, the 5-parameter NRTL model25 was used to predict the bubble pressure for a given temperature and liquid composition. The NRTL equation can be represented as follows:
,where
,
, and
Aji = uji - uii.
For isothermal data sets, binary interaction parameters are considered to be composition-dependent:
.
For isobaric data sets, temperature dependence of the parameters is represented as follows:
.
For a complete T-p-x-y data set, five parameters are determined 
. After completion of the fitting process, the following criteria are applied:
where N is the number of properties values, the superscript exp indicates experimental data, and the superscript cal indicates values calculated with the NRTL equation. If Δp and Δy are less than 1, the data set passes the test.
The quality factor for the Van Ness Test is calculated as follows:
Ftest2 = 2/(Δp + Δy) with the limits 1 ≤ Δp ≤ 10, 1 ≤ Δy ≤ 10.
Van Ness Test Results: Output screen details
A more comprehensive modeling test for isobaric data sets can be performed with Barker's method26 that incorporates experimental excess enthalpy data. Implementation of Barker's method planned as a future extension of this work.