DBPapers
DOI: 10.5593/sgem2017H/15/S06.035

IS USING RIGOROUS THERMODYNAMICS FOR FULLY COMPOSITIONAL RESERVOIR DEPLETION SIMULATION WORTH THE EFFORT?

V. Gaganis, N. Varotsis
Thursday 23 November 2017 by Libadmin2017

References: 17th International Multidisciplinary Scientific GeoConference SGEM 2017, www.sgemviennagreen.org, SGEM2017 Vienna GREEN Conference Proceedings, ISBN 978-619-7408-26-3 / ISSN 1314-2704, 27 - 29 November, 2017, Vol. 17, Issue 15, 273-280 pp; DOI: 10.5593/sgem2017H/15/S06.035

ABSTRACT

Compositional reservoir simulation runs involve, among others, the solution of the phase behavior of the hydrocarbon fluid phases at each block and at each time step. This iterative, high CPU time demanding calculation renders history matching and production optimization as extremely time consuming tasks.
This paper presents a review of the available methods for speeding up the conventional phase behavior computations during compositional simulation Emphasis is given to the fluid-specific soft computing models which can be prepared offline to provide values of the fluid properties of interest by means of simple, non-iterative calculations. These models take advantage of special features of the phase behavior calculations in a flow simulation and provide very rapid answers during simulation. Such methods include, among others, proxy models for treating both the phase stability and phase split problems using machine learning tools and use of soft computing techniques to develop pressure-explicit interpolation schemes for the direct, non-iterative calculation of fluid specific equilibrium coefficients values.
This work discusses the pros and cons of the different approaches and it presents, for a selected reservoir model, the accuracy and CPU requirements of the respective simulation runs. The results obtained show that using rigorous thermodynamics during compositional reservoir simulation is an expensive luxury since the utilization of smart soft computing approaches make practically no compromise to accuracy whereas they can reduce the CPU requirements down to at least one order of magnitude.

Keywords: compositional reservoir simulation, phase behavior calculations, algorithm acceleration