Significant compositional changes have been recently documented in laboratory analyses of remaining oil in core samples from mature steamfloods. In order to properly account for these changes, the production of hydrocarbon condensate (casing blow) from heavy oil reservoirs, and the formation of an in-situ solvent bank, a fully compositional, equation-of-state approach in thermal reservoir simulation is needed. Dead oil thermal simulation will not account for these recovery mechanisms.
A survey of equation-of-state development for lighter oils reveals that traditional methodologies using a C7+ heavy split with a gamma distribution and α=1 may not be correct for some medium to heavy oils. Extensive gas chromatography (GC) analyses reveals a "double hump" or bimodal compositional distribution in some medium to heavy oils. This bimodal distribution is often overshadowed in lighter oils due to increased presence of C1-C6 components. Because of this bimodal distribution and absence of significant C1-C6 composition, a different methodology and series of laboratory test must be used to develop equation-of-state parameters as well as account for complex interactions between polar water molecules and non-polar hydrocarbon molecules which occur in the process of steamflooding. This data is then used to tune equation-of-state parameters for use in compositional thermal simulation. This new technique is presented as an alternative model for EOS parameter development for medium to heavy oils and is more economical than previous methods which require a full ASTM-2892 distillation. San Ardo crude oil (120API) is used to demonstrate the technique. The full 40 component model can then be pseudoized into fewer components for field scale simulation.