PROPERTIES OF RESERVOIR GASES - Petroleum Reservoir Fluid Property Correlations - page 22

Table 2-5. Relative errors to published gas viscosity correlations compared to data described in
table 2-4
Viscosity correlation* Average relative error, % Average absolute relative error, %
Lee et al. (1966)* -1.60 2.26
Londono et al. (2005)* -2.66 3.08
Sutton (2007)* 2.05 3.10
Poling et al. (2001)* -0.61 3.34
Lee et al. (1964)* -3.23 3.70
Carr et al. (1954) † -9.81 9.81
* Piper et al. correlation was used to estimate pseudocritical properties and Dranchuk-Abou-Kassem equation was used to estimate z-factor
for density calculations as needed.

† Standing and Dempsey equations represented the Carr, Kobayashi, and Burrows graphs.

Figures 2-5 through 2-7 show the results of the best four correlations,
with the data set sorted and divided into approximately equal subsets
based on temperature, pressure, and gas specific gravity.
The data were also sorted and split by gas viscosity. The results looked
very much like figure 2-6.
These figures show that the Lee, Gonzalez, and Eakin (LGE) equations
give the best results with this set of data. Londono et al. reported the
same results.
The correlations look fairly good across these ranges of temperatures,
pressures, and gas specific gravities. However, the maximum pressure
of the data set is well below pressures often encountered in petroleum
reservoirs. Also, there is no way of determining how well the correlations
predict gas viscosities for gas condensates. Calculations with the LGE
equations were extended to very high pressures and high specific
gravities. No reservoir gas data exist to compare with the results, but the
calculated values were well behaved and looked reasonable.