Laboratory Tests on Heavy Oil Recovery by Steam Injection

This paper presents results of laboratory experiments conducted to determine the effect of various parameters on residual oil saturation from steamdrives of heavy-oil reservoirs. These experiments indicated that remaining oil saturation, both at steam breakthrough and after passage of several PV of steam, is a function of oil/water viscosity ratio at saturated steam conditions. Introduction Considerable attention has been given to thermal techniques for stimulating production of underground hydrocarbons, particularly the more viscous oils production of underground hydrocarbons, particularly the more viscous oils and tars. Steam injection has been studied as one means of heating oil in place, reducing its viscosity, and thus making its displacement easier. place, reducing its viscosity, and thus making its displacement easier. A number of investigators have measured residual oil saturations remaining in the steam zone. Willman et al. also analyzed the steam displacement process to account for the oil recoveries observed. A number of methods have been developed to calculate the size of the steam zone and to predict oil recoveries by application of Buckley-Leverett theory, including the use of numerical simulation. The work described here was devoted to an experimental determination of oil recovery by steam injection in linear systems. The experiments were unscaled as far as fluid flow rates, gravity forces, and heat losses were concerned. Part of the study was to determine recoveries of naturally occurring very viscous tars in a suite of cores containing their original oil saturation. The cores numbered 95, 140, and 143 are a part of this group. Heterogeneities in these cores, however, led to the extension of the work to more uniform systems, such as sandpacks and Dalton sandstone cores. Our interest was in obtaining an overall view of important variables that affected recovery. In particular, because of the significant effect of steam distillation, most of the oils used in this study were chosen to avoid this factor. We also studied the effect of pore size on the residual oil saturation. As part of this work, we investigated the effect of the amount of water flushed through the system ahead of the steam front in several ways:the production rate was varied by a factor of four,the initial oil saturation was varied by a factor of two, andthe rate of heat loss was varied by removing the heat insulation from the flow system. Description of Apparatus and Experimental Technique Two types of systems were studied: unconsolidated sand and consolidated sandstone. The former type was provided by packing a section of pipe with 50–70 mesh Ottawa sand. Most runs on this type of system were in an 18-in. (45.72-cm) section of 1 1/2 -in. (3.8 1 -cm) diameter pipe, although runs on 6-in. (15.24-cm) and 5-ft (152.4-cm) lengths were also included. Consolidated cores 9 to 13 in. (22.86 to 33.02 cm) long and approximately 2 1/4 in. (5.72 cm) in diameter were sealed in a piece of metal pipe by means of an Epon/sand mixture. A photograph of two 9-in. (22.86-cm) consolidated natural cores (marked 95 and 143) from southwest Missouri, containing original oil, is shown as Fig. 1. In all steamdrive runs, the core was thermally insulated to reduce heat loss, unless the effect of heat loss was specifically being studied. Flow was usually horizontal except for the runs in which the effects of flushing water volume and of unconsolidated-sand pore size were examined. Micalex end pieces were used on the inlet end in initial experiments with consolidated cores to reduce heat leakage from the steam line to the metal jacket on the outside of the core. During most runs, however, the entire input assembly eventually became hot. SPEJ p. 417