Evaluation of Efficiency Gas EOR Techniques Using Advanced PVT-model

Abstract

Conventional way of building PVT-models of reservoir fluids consists of creating the simplistic so-called black-oil model. Results from PVT laboratory (fluid densities, bubble-point pressure, GOR, etc…) are using as input data for PVT correlations. All other properties of the reservoir fluids are calculated from widely known PVT correlations. But simplistic black-oil models were not designed for the modelling of change of fluid state (oil to gas and vice versa). Also black-oil models are not capable to simulate effects on the oil-gas contact. In order to overcome above mentioned issues, compositional PVT-models have been introduced. These models solve equation of state (EOS) for each component. They add number of equations to the numerical simulator (lumping procedure could be introduced), but allow to accurately simulate effects on the gas-oil contact and transfet of one hydraucarbon phase to another. But usually a few results of fluid analysis are availbale and the question is: how to use all those results to build a non contradictive PVT model which is in line with all PVT experiments results? In order to overcome this challenge, the workflow for quality control (QC) and analysis of fluid PVT data is proposed. That workflow was used to analyse data from reservoirs an Abu Dhabi Offshore field. After selection and cleaning of data set, the lumping of initially broad composition (C36+) was executed. Then regression algorithms from commercially available software have been used to match lumped composition properties with results of PVT experiments. Results of PVT experiments from other wells were used as a blind test for derived composition.