Gravity drainage can result in a high recovery efficiency for high permeability oil reservoirs with low viscosity oil and low interfacial tension. In order to make reliable field wide performance predictions of this production mechanism, it is necessary to obtain accurate determination of gas-oil relative permeabilities and residual oil saturation. Gas-oil relative permeabilities been extensively investigated, performing:–displacement experiments in various conditions (differential pressure, interfacial tension, connate water or not, short or long core, room or reservoir conditions).–dynamic capillary desorption experiments. Special emphasis has been given to the actual physical meaning of the residual oil saturation and the way it is reached. The main results are:–Gas-oil displacement experiments performed in laboratory conditions (high interfacial tension) with various pressure drops give the same relative permeability if interpreted taking capillary effect into account. The final experimental oil saturation, which depends on the maximum acceptable injection pressure (turbulent flow problems) is different from the "true" ROS (corresponding to kro = 0).–Capillary desorption experiments in lab conditions with numerical interpretation of the transient production, may be used to determine ROS and oil relative permeability at high gas saturation.–Interfacial tension (in the range 0.6 - 30 mN/m) has little influence on the relative permeabilities.–Presence of connate water has little influence: oil and water behave as a single liquid phase.–Relative permeabilities obtained from laboratory conditions (short core) and reservoir conditions (long core) experiments are very close.–The final desaturation (ROS around 5 %PV) is reached very slowly, due to the very low level of oil relative permeability (about 10-5). A reliable method to determine relative permeabilities in laboratory conditions has been established, by performing both displacement and capillary desorption experiments. Introduction It is well known that the gravity forces play an important role in the oil recovery by the so called gravity drainage process, which occurs in permeable reservoirs, saturated with light oil displaced by gas. P. 339^

In order to properly account for wettability and oil trapping, laboratory waterflood experiments need to be carried out with reservoir fluids and at reservoir flooding velocities. An interpretation procedure has been defined for such measurements with the help of phenomenological experiments conducted under laboratory conditions with an apparatus allowing direct measurements of saturation and pressure in each phase at different points in the core. Experiments can be interpreted with no ambiguity. They indicate that relative permeabilities are independent of flow rate except near the residual oil saturation. Capillary pressures however depend on flow rate and porous medium wettability:–in some porous media (mainly sandstones), the imbibition capillary pressure always remains positive and changes very little with the flow rate;–in other media (mainly carbonates) the capillary pressure, which is generally positive during the initial oil drainage phase, becomes negative immediatly behind the water front. The higher the waterflooding rate the more negative the capillary pressure. This make sit impossible to correctly interpret water-flooding experiments using static capillary pressure data. As direct measurement of dynamic capillary pressure is not feasible for experiments carried out pressure is not feasible for experiments carried out under reservoir conditions, a procedure, based on the preceding conclusions was developed to interpret such preceding conclusions was developed to interpret such experiments. By combining the results of experiments at high velocity and at low velocity it is possible to determine relative permeabilities, residual oil saturation, dynamic capillary pressure and accordingly wettability. Introduction Temperature, nature of oil, flooding velocity, microscopic distribution of fluids affect the displacement of oil by water as shown by several authors (ref. 1 to 11). Therefore, experiments carried out under reservoir conditions are those offering the best representativity guarantees. However, as low flooding velocities are generally encountered in reservoirs, gravity, viscous and surface (capillary) forces cannot be neglected. Besides, information obtained is necessarily limited by the severe conditions under which experiments are curried out and concerns only - produced oil and water, - pressure drop across the core (underexperimental conditions with constant injection rate). Such partial information cannot allow a complete interpretation of experiments i.e. determination of relative permeability and imbibition capillary pressure curves. pressure curves. We, then, dedicated ourselves to the definition of a procedure to fully interpret experiments under reservoir conditions; a better understanding of all concerned phenomena was necessary. For that purpose, we used an experimental device working under laboratory conditions and with which flooding phenomenology could be studied through pressure and saturation measurements made respectively pressure and saturation measurements made respectively in each phase and at different points in the core. Such device (alpha type) provides sufficient information for full experiment interpretation (simultaneous determination of relative permeability and capillary pressure curves) without it being necessary to get pressure curves) without it being necessary to get information from other experiments. All elements are absolutely certain, though they are not quite representative of reservoir, as they are obtained under laboratory conditions; various experiments on the same medium carried out using such device are therefore comparable.

Reservoir engineers have many problems in forecasting the oilfields exploitation if they use the laboratory data on the relative permeability curve vs saturation. permeability curve vs saturation. This leads one to wonder whether this concept is valid in transient flow, whether data have been made by suitable methods, whether the static relation between capillary pressure vs saturation is preserved in variable flow, whether lack of familiarity with combination laws concerning this function is not the reason for the dissatisfaction of the engineers. The present paper deals with the first three points. present paper deals with the first three points. Three series of oil displacement by gas have been made on a homogeneous sandstone core. Pressure in the two phases and saturation Pressure in the two phases and saturation (gammagraphy) have been measured all along the core. Pressure profiles have been found by the least-square method. Relative permeability has been calculated on each point and at each time by continuous equations. These studies have shown us that the relation relative permeability vs saturation is the same all along the transient fluid flow and that the relation capillary pressure vs saturation is the same as the one established by the restored states static method. Thus, the validity of this concept concerning transient fluid flow is proved. The Welge, Johnson, Bossler and Naumann method arrives at different results from those we have obtained. To explain this difference we show that the theoretical hypotheses of their method are far removed from reality. In conclusion, experimental results are simulated by numerical model taking into account capillary phenomena. The limit conditions established by experiments and used in numerical simulation are (1) inflow side (constant gas-injection pressure; pressure gradient is zero in oil) and (2) outflow side (out-gas pressure equal to the displacement pressure; pressure in oil is zero). The pressures are relative pressures. pressures Introduction When sweep efficiency of one fluid displacing another has to be predicted, the approach to the problem is nearly always identical. Laboratory displacement tests are performed on small samples that are supposed to performed on small samples that are supposed to be representative of the reservoir. After analyzing the data by the simplified method of Welge, Johnson, Bossler and Neumann, the relative permeability-saturation relationships are derived. In order to characterize more completely the porous medium, the capillary pressure-saturation relationship is also pressure-saturation relationship is also determined by the restored-state technique.