ef6b01419_si_001.pdf (179.07 kB)
Download filePrediction of Viscosity for Characterized Oils and Their Fractions Using the Expanded Fluid Model
journal contribution
posted on 2016-08-09, 00:00 authored by F. Ramos-Pallares, S. D. Taylor, M. A. Satyro, R. A. Marriott, H. W. YarrantonA methodology has been developed
to predict the viscosity of crude
oils and their fractions from a distillation-based oil characterization.
The maltenes were characterized as a set of pseudo-components with
properties determined from established generalized correlations. The
asphaltene fraction was characterized as a single component and its
properties were measured. The viscosities of the pseudo-components,
asphaltenes, whole oils, and their fractions were determined with
the Expanded Fluid (EF) viscosity model. The inputs for the model
are the pressure, the density of the fluid at a given pressure and
temperature, the dilute gas viscosity calculated from established
generalized correlations, and three fluid-specific parameters: c2, ρs°, and c3.
Densities were calculated using the modified Rackett correlation with
the Tait-COSTALD compressibility correction. The c3 parameter was determined from a previously developed
correlation. New correlations were developed for the c2 and ρs° parameters of the maltene pseudo-components,
as a function of their boiling point and specific gravity. The parameters
for the asphaltene fraction were estimated based on the measured viscosity
of molten asphaltenes. The EF parameters for the whole oil, or any
oil fraction, were determined with mass-based mixing rules and binary
interaction parameters, calculated from a previously developed correlation.
To develop and test the proposed approach, density and viscosity data
were collected for 40 distillation cuts from 6 oils, 7 maltenes, 2
asphaltenes, 3 partially deasphalted oils, and 14 dead oils. Using
this model to predict crude oil viscosity under any conditions requires
the distillation assay data, the asphaltene mass content, and the
specific gravity and molecular weight of the oil. The approach was
tested on a development and test dataset of 4 crude oils (this study)
and an independent test dataset of 4 oils from the literature with
overall average absolute relative deviation (AARD) values of 41% and
43%, respectively. Single multiplier tuning of the c2 parameter to one viscosity data point halved the error.
Tuning both the c2 and ρs° parameters
using two viscosity data points reduced the AARD to <8% in both
cases.