A STATISTICAL METHOD FOR ACCEPTANCE OF CRUDE OIL VISCOSITY-TEMPERATURE CURVES

J. C. M. Dias, P. F. Aguiar

Abstract


The aim of this work is to present a statistical model for acceptance tests of viscosity-temperature curves of crude oils, to help identifying potential deviations caused by experimental errors, inadequate methods or even oil behavior anomalies. An empiric exponential function, which is very used in the petroleum industry to determine relationships between viscosity and temperature regarding oil and by-products, was linearized to serve as model for regression of experimental values obtained from several oil samples. Through analysis of the regression model variance, we can determine the minimum significance level so that experimental values acquired for each oil sample can be in accordance with a theoretical model. As a result, two acceptance criteria were established: one for using data under measurement conditions, and another to be used in estimate calculations, such as interpolations and extrapolations.

Keywords


ANOVA; chemometrics; crude oil; viscosity

Full Text:

PDF


DOI: http://dx.doi.org/10.5419/bjpg2011-0003