The Role of Dispersion in Enhanced Gas Recovery and Gas Field Pressure Maintenance

Enhanced Gas Recovery (EGR) is the process whereby an inert gas like nitrogen or flue gas is injected in a gas reservoir to improve hydrocarbon gas recovery. One of the objectives of EGR is recovery of remaining gas in place at the prevailing abandonment pressure by sweeping native hydrocarbon gas with an inert gas. This paper treats the reservoir engineering aspects of dispersion in gas displacement by nitrogen. Relevant theory and knowledge from literature are applied to an example sandstone gas reservoir. The displacement is typically miscible, and the higher viscosity and density of the injected nitrogen over the native hydrocarbon gas improves the stability of the vertical displacement front. However, dispersion in the reservoir is another potential source of spreading of the front. This leads to early nitrogen breakthrough and a slowly growing nitrogen concentration in the production stream that needs to be dealt with prior to sales through N2 removal or dilution of the produced gas with other gas streams. Reservoirs with low formation dispersivity are therefore the most suitable targets for EGR. This leads to the selection of homogeneous reservoirs with short correlation distances of depositional features. Formation dispersivity is ideally measured upfront using a tracer push-pull test. As a result of the physics of the dispersion process a line drive with a large displacement well spacing provides an optimum selection as (horizontal or vertical) well configuration. Selection of high viscosity injection gas helps to increase the stability of the displacement front. Stabilization of the injection front by foam would significantly enlarge the targeted group of fields for EGR to include reservoirs with more adverse heterogeneity. R&D is required to establish a likely reduction in dispersion. Accurate modelling of the mixing process is possible by tagging the injection fluid with a passive tracer while solving the advection equations explicitly using a higher order scheme to reduce numerical dispersion. Only physical dispersion at the sub-grid scale should be included. This modelling method could however lead to unstable displacement in the simulator because the density and viscosity contrasts are ignored.