Comparative Evaluation of a Functions and Volume-Translation Strategies to Predict Densities for Gas(es)-Heavy Oil/Bitumen Systems

In this work, a unified, consistent, and efficient framework has been proposed to better predict the density of a gas(es)-heavy oil/bitumen system by using the Peng-Robinson equation of state (PR EOS) and Soave-Redlich-Kwong (SRK) EOS together with α functions and volume-translation (VT) strategies, respectively. With a database comprising 218 experimentally measured densities for gas(es)-heavy oil/bitumen systems, five α functions defined at a reduced temperature (Tr) of 0.70 as well as three new α functions at Tr = 0.60 together with four VT strategies are selected and evaluated. For α Functions 1 to 4 defined at Tr = 0.70, VTs 1 to 4 lead to an overall absolute average relative deviation (AARD) of 7.21%, 9.74%, 7.02%, and 7.16%, respectively, for predicting the mixture densities. For α Function 5 defined at Tr = 0.70, these four VT strategies predict the mixture density with an AARD of 3.13%, 5.01%, 2.92%, and 2.56%, respectively. As for the two new α Functions 7 and 8 defined at Tr = 0.60, these four VT strategies predict the mixture density with an AARD of 1.38%, 2.57%, 1.34%, and 1.67%, respectively, among which VT 3 has a very close prediction compared to an AARD of 1.31% obtained from the ideal mixing rule with effective density (IM-E).