TITLE:
Data Consistency Tests through the Use of Neural Networks and Virial Equation. Application of the Proposed Methodology to Critical Study of Density Data
AUTHORS:
Abdeslam Hassen Meniai, Serge Laugier, Hakim Madani, Dominique Richon
KEYWORDS:
Consistency Tests, Hexafluoropropylene, Neural Networks, Vibrating Tube Densimeter, Virial Equation
JOURNAL NAME:
Open Journal of Physical Chemistry,
Vol.1 No.3,
November
25,
2011
ABSTRACT: This paper focuses on a very important point which consists in evaluating experimental data prior to their use for chemical process designs. Hexafluoropropylene P, ρ, T data measured at 11 temperatures from 263 to 362 K and at pressures up to 10 MPa have been examined through a consistency test presented herein and based on the use of a methodology implying both neural networks and Virial equation. Such a methodology appears as very powerful to identify erroneous data and could be conveniently handled for quick checks of databases previously to modeling through classical thermodynamic models and equations of state. As an application to liquid and vapor phase densities of hexafluoropropylene, a more reliable database is provided after removing out layer data.