Oil & Gas - Maximising Recovery Programme (formerly SHARP) IOR Views

Introduction and Aim

The aim of this project is to model real petroleum oils using mixtures of Paraffin, Naphthene and Aromatic (PNA) components. Since PNA components can straddle the full range of properties of real oils it should be possible to map the properties of a real oil onto the properties of a simple mixture of PNA components. Once such a mapping is in place computation can be performed on the model system, which, due to the mapping, will mimic the properties of the real system. The principal advantage of this process will be that the computations can be checked against experiments on the model system, which is not possible with current lumping schemes, based on boiling fractions. This proposal focuses on carrying out the mapping based on phase densities.

Background

This proposal is a logical progression from a series of research projects carried out at the University of Liverpool. Initially, we worked on understanding the interfacial tension behaviour of nitrogen-hydrocarbon systems in relation to Improved Oil Recovery (IOR) by nitrogen injection, as interfacial tension is an important parameter for many reservoir-engineering assessments. As part of this work we began developing a fundamental statistical mechanical technique to predict the interfacial tension. This work showed that prediction was possible but further work was necessary to apply it to real fluids. Recently, effort has been expended in speeding up the computational procedures to assist in the eventual task of computing the interfacial tension of multi-component hydrocarbon mixtures. However, even with such a tool available the obstacle to fully implementing it will be how to select the appropriate input information. Of all the thousands of components in a real oil how do we choose the most appropriate ones? The standard lumping schemes are inadequate because they join in one pseudo-component, molecular types having opposing interfacial tension behaviour. This led us to consider the types of molecules in petroleum fluids and their relative contribution to the interfacial tension. We identified intra- and inter-pseudo-components and how mixtures of these might more precisely reflect the physical properties associated with the interfacial tension.

Reservoir fluids are composed of a very large number of components most of which are unresolved by standard analysis techniques. Currently, because of the complexity of reservoir simulations, these complex liquids must be reduced to a small number of pseudo-components (say, twelve for compositional simulation), if the simulations are to obtain solutions in acceptable time spans. Coupled with this, only minimal compositional analysis of the light components of the reservoir fluids, a true boiling curve, and the gas-oil ratio are normally commercially available. Occasionally, as an addition, an estimate of the paraffins, naphthenes and aromatics (PNA analysis) together with the density of the individual cuts of the true boiling fractions are given. Even so, the current best standard analysis procedures would only distinguish about 100 components from the many thousands in a reservoir fluid.

Empirical methods are utilised for estimating many of the parameters used in equations that describe the physical properties of the reservoir fluid components. Lack of sound theoretically based methods for equation parameterisation means that extrapolation to more extreme conditions become unreliable, with resultant poor prediction of the actual physical properties under reservoir conditions. For the foreseeable future, the physical properties in reservoir simulation will be described by a small number of pseudo-components. The fluid properties computed from these groupings depend heavily on which empirical pseudoisation technique is used. Furthermore, these groupings generally arise from gas chromatographic analysis techniques and hence are based on boiling point. Consequently, this way of selecting pseudo-components will be appropriate only for physical properties related to boiling points. This clearly shows the need for improved methods of fluid characterisation.

Reservoir entity simulations play an important role in the evaluation and management of hydrocarbon reservoirs. For condensate reservoirs, physical properties undergo significantly greater changes in magnitude than in black oil reservoirs, and thus the physical properties are more important in these cases. Physical property predictions play a major role in these simulations and may take up to 80 per cent of the computing time. Thus, improved accuracy in these properties must be weighed against the potential increase in program execution time. Improving computing speeds give rise to a number of options. Obviously, solutions to current problems can be achieved more quickly. Alternatively, the potential exists to simulate reservoirs of increasing complexity. With condensate reservoirs, it is also possible to increase the number of components in the simulation to obtain a better physical property description. This combined with better mapping schemes may result in improved fluid property predictions.

Deliverables

The deliverables from this project would be:

• Effective fluid characterisation of selected sponsor's petroleum oils.
• Phase densities of the selected PNA mixtures.
• A best practice of improved pseudoisation of reservoir fluids with respect to phase densities.
• Validation of mapping procedure for phase density.

Partners and Funding

This proposal already has the potential support of the DTI but in order for the project to fly it requires three additional oil company sponsors each contributing £20k/year for two years. Anyone interested please contact Dr John Satherley (js1@liverpool.ac.uk).