Published on: 12/13/2021
To understand how fluid behaves is fundamental for exploration and production in oil and gas fields. Reservoir engineers rely on PVT relations to calculate the size of oil and gas reserves, the volume and timeframe for production, and the efficiency of enhanced oil recovery (EOR) methods. Production engineers use phase behavior data to design surface facilities and pipelines, to calculate the multiphase flow inside pipes, and to prevent flow assurance issues that might appear. Such calculations are made over a range of temperatures from the surface to reservoir conditions [1].
In general, experimental data available to describe the fluid produced is composed of the fluid composition, typical PVT experiments (as discussed in this blog post), and viscosity values for a specific range of pressures and temperatures. However, this data is too restricted to meet the needs of the oil & gas industry. To conduct reservoir and production analyses, engineers also need iInformation about phase equilibrium and other physical, thermal, and transport properties of each phase, such as thermal conductivity and surface tension, within a wide range of pressures and temperatures. .
Therefore, characterizing the fluid and accurately estimating its phase behavior and properties through simulation tools is a key part of the workflow needed to define recovery strategies, prevent flow assurance issues, and design pipelines and surface facilities.
To help engineers obtain a reliable representation of fluid behavior and its properties, ESSS is developing a Fluid Analysis and Simulation Environment (FASE) software. This software performs fluid characterization based on an Equation of State (EoS) modeling, simulating the phase equilibria of complex hydrocarbon mixtures and estimating their physical, thermodynamic, and transport properties.
Two types of cubic Equations of State are available in FASE: Peng-Robinson (PR-78) and Soave-Redlich-Kwong (SRK). Both models can be used with Peneloux volume correction in order to improve the predictions of volumetric properties of the liquid phase. Default binary interaction parameters (BIPs) for the classical van der Waals one-fluid mixing rules are provided, but custom values can be specified as well.
In addition, Huron-Vidal mixing rules can be used in combination with classical mixing rules to achieve adequate representations of polar interactions. The models are implemented under the residual Helmholtz free energy framework, and all derivatives are computed using automatic differentiation [2]. This approach circumvents errors caused by wrong manual derivations of complex analytical expressions or code implementation, without losing numerical precision.
Pedersen’s characterization method [3] is used to estimate the molar fractions and critical properties of the heptane-plus fractions. To reduce the computational cost of the simulations, light or heavier components can be grouped into pseudo-components using mass weighted averages.
When lumping heavier components, it’s only necessary to input the number of desired pseudo-components, and FASE decides automatically which intervals of single carbon numbers are grouped together using a criteria based on mass balance among pseudo-components.
FASE software can solve phase equilibria problems through flash calculations for up to four coexisting phases. Four modes of operations are available: custom mode, where the user provides a list of conditions of pressure and temperature; fixed temperature, where isothermal calculations are performed within a given pressure range; fixed pressure; and grid mode, in which pressure or temperature ranges are specified.
Phase properties are automatically computed in the Flash module. At the end of the simulation, FASE provides the values of the following phase properties: molar fraction, molar volume, weight fraction, volume fraction, density, molecular weight, viscosity, thermal conductivity, interfacial tension, enthalpy, internal energy, entropy, heat capacities (Cp and Cv), Joule-Thomson coefficient, and velocity of sound.
Beyond these capabilities, properties from typical PVT experiments (CCE, DL, and Separator Tests) can also be calculated. When comparing these PVT simulations with experimental data, it can be seen that, although the characterization procedure provides good estimates for both composition and properties of the mixture, the heavier components’ contribution still represents the major source of errors. Thus, when PVT data is available, FASE can adjust the model to match experimental values through robust regression algorithms in order to obtain a representative model of a given fluid. The Peneloux volume shift constant, binary interaction parameters, and coefficients of correlations used to estimate the properties of the heavier fraction are tuned using a multi-stage optimization procedure.
Get more information about FASE and ESSS’s other simulation software solutions and contact us.
[1] Whitson, Curtis H., and Michael R. Brulé. Phase behavior. Vol. 20. Richardson, TX: Henry L. Doherty Memorial Fund of AIME, Society of Petroleum Engineers, 2000.
[2] Leal, A.M.M. (2018). autodiff, a modern, fast and expressive C++ library for automatic differentiation. Available at: https://autodiff.github.io (Last access: August 3, 2021).
[3] Pedersen, K. S., Christensen, P. L., Shaikh, J. A., & Christensen, P. L. (2006). Phase behavior of petroleum reservoir fluids. CRC press.
