Effects of Experimental Errors on Thermodynamic Consistency and on Representation of Vapor-Liquid Equilibrium Data

Abstract

A thermodynamic consistency test based on the "local area" test is developed which permits analysis of vapor-liquid equilibrium data in terms of measurement uncertainties in the independent variables. The effect of random measurement errors on the consistency test and on the uncertainty in the activity coefficients is illustrated. Uncertainty in the measurement of liquid composition may cause large uncertainty in the activity coefficients but be difficult to detect by means of consistency test. Comparing the test results to the null value is shown to be of less value than comparison with the bounds of a confidence region.

Vapor‐Liquid equilibrium: Part I. An appraisal of data reduction methods

“…Although the vapor pressures PISat and PgSat are functions of temperature, fixed values are used with the entire data set, and random error is not a consideration. Estimates of the standard deviations in the functions g/xlx2 and In(yl/y2) when these quantities are calculated for a finite data set are given by equations derived in the manner of Ulrichson and Stevenson (1972). For g/xlx2:…”

Reduction of thermodynamic data by means of the multiresponse maximum likelihood principle

“…His criterion is based on the assumption that the covariance matrix is constant. This is not true, even if the variances of the direct measurements (x,T,y and P) are constant (Van Ness et al, 1973;Ulrichson and Stevenson, 1972. ) Furthermore, the Schubert criterion contains four dependent variables instead of two.…”

Method of evaluation and reduction of vapor‐liquid equilibrium data of binary mixtures

“…Evaluation of experimental data has been made traditionally by tests of consistency. However, recent papers have questioned the former use of the area test; Van Ness et al (1973) recommend concentration of efforts in improving measurements rather than obtaining redundant data for such tests; Samuels et al (1972) and Ulrichson and Stevenson (1972) redirect the attention towards the effect of random errors, incorporating the experimental evaluation of these errors into the use of a test.…”