A Robust Workflow for Reliably Describing Reservoir Fluid PVT Properties Using Equation of State Models

Abstract

AbstractOne of the key elements in modern reservoir exploration and management is describing reservoir fluid phase behavior and physical properties commonly referred to as pressure-volume-temperature, or PVT data. Typically, PVT data come from laboratory tests, empirical correlations, and Equation of State (EOS) models. It is common practice to describe the phase behavior or PVT data through EOS models tuned to laboratory measurements on reservoir fluid samples. After a sample is received at a laboratory, a portfolio of PVT laboratory tests are performed. The results are quality checked and the appropriate data are selected to tune an EOS model to obtain an accurate EOS description of the reservoir fluid. Each step in this process requires judgment and decisions from the corresponding domain experts to achieve physically sound PVT relations and calculate the required properties. Such EOS based modelling processes are time consuming, expensive and exposed to various risks due to multiple human interventions.In the work presented in this paper, a study was conducted to explore the feasibility of a defined step EOS based modelling workflow using a limited, but defined, laboratory data set as the basis for characterizing and tuning the EOS model. In the development of a standardized workflow, a variety of EOS characterization and tuning methods were established to accommodate the diverse and complex nature of reservoir fluids. The methods considered include a modified Pedersen's method, a gamma distribution based method, and two methods based on single carbon number (SCN) composition and aromaticity factors. Despite their differences, the methods follow the principal objectives to be operator independent, robust, thermodynamically consistent, and numerically simple. Apart from the fluid composition, the only PVT data required for the proposed workflow were the saturation pressure, densities and the volumetric data obtained from constant composition expansion (CCE) measurements. The CCE test is advantageous because it is non-destructive to the sample being tested and can be performed quickly and reliably either in the laboratory or at a well-site. An optimized EOS model utilizing appropriate fluid characterization and tuning method is then selected based on a pre-defined Key Performance Indicator (KPI) derived from the deviation of the model predictions with the experimental data. With this optimal model defined, all other PVT data, such as those from sample destructive differential liberation (DL) or constant volume depletion (CVD) tests, can be reliably predicted.The workflow was validated using extensive PVT data measured for a variety of reservoir fluids including black oils, volatile oils, and gas condensates. The results have shown that the proposed workflow can reliably model PVT phase behavior and properties for the majority of tested reservoir fluids. This article will thoroughly discuss the details of this workflow and the modelling results on four benchmark fluids of different types.