The Effect of Microscopic Core Heterogeneity on Miscible Flood Residual Oil Saturation

Abstract

Much effort has been devoted to understand the mechanisms responsible for the variation of CO2 flood residual oil saturations found both in the laboratory and in the field. Some of the many possible explanations are the detailed nature of CO2 possible explanations are the detailed nature of CO2 /oil phase behavior, trapping of oil by mobile water, viscous fingering, and bypassing of oil due to the micro-pore structure of a given porous medium. The purpose of this study was to investigate the influence purpose of this study was to investigate the influence of rock characteristics alone on miscible displacement behavior through a combination of displacement testing and modeling. Modeling was used to characterize the displacements and to extrapolate the observed phenomena to untestable conditions. The displacement phenomena to untestable conditions. The displacement tests were used to calibrate the model and to test the model's predictive capabilities. A number of stabilized CO2 displacements and xylene displacing iso-octane tests were conducted in both outcrop sandstones and San Andres reservoir carbonate core samples. By flooding with xylene displacing iso-octane, phase behavior complications and viscous fingering were avoided. Such an idealized displacement represented the "best case" laboratory miscible displacement. Effluent concentration profiles for the xylene displacing iso-octane floods profiles for the xylene displacing iso-octane floods were matched to the capacitance-dispersion model of Coats and Smith to estimate the magnitude of axial dispersion and so-called "dead-end" pore volume or capacitance. The latter quantity was found to be essentially zero in outcrop sandstones but quite variable in the highly heterogeneous San Andres carbonates, attaining values as high as 50%. Axial dispersion followed similar trends. The carbonates displayed more variable and larger axial dispersion than the sandstones. First contact and CO2 miscible flooding residual oil saturations were compared in the same core samples. Both measures of residual oil saturation were in agreement and depended on the particular sample. These data suggest microscopic heterogeneity is a primary determinant of residual oil saturation to miscible flooding when viscous fingering is controlled. Microscopic heterogeneity resulting from a non-uniform pore structure was characterized by the level of dispersion and capacitance in the rock sample. In addition, both laboratory data and model prediction showed that the effect of dispersion and capacitance on residual oil saturation decreased as the displacement was extrapolated to reservoir rates and system lengths of a few feet. This implies that microscopic heterogeneity is more important in laboratory systems than in field displacements. Introduction Carbon dioxide miscible flooding is one of the most promising tertiary recovery techniques. Considering the expense and complexity of the process, applicability to a given candidate reservoir process, applicability to a given candidate reservoir can, at this time, only be determined by a thorough laboratory, numerical and economic analysis. Usually, the results of such an evaluation depend heavily on an estimate of the CO2 unit displacement efficiency. The unit displacement efficiency is measured by field pilot testing or from laboratory core displacement tests in which a laboratory CO2 -flood residual-oil saturation, S, is compared to a waterflood residual-oil saturation, S. A wide variation in CO2 S has been observed in both field pilot tests and laboratory displacements conducted above the multi-contact miscibility pressure. This work was undertaken to further understand the reasons for this variation. A better understanding of the mechanisms responsible for the variability of S in the laboratory would guide interpretation of field pilot data and possibly eliminate the need for extensive field pilot testing to obtain unit displacement efficiency. Variability of S from laboratory data has been attributed to such factors as CO2 /oil phase behavior, rock tore structure trapping of oil by mobile water, flooding rate, and rock wettability as well as experimental technique.

New Method of Incorporating Immobile and Nonvaporizing Residual Oil Saturation into Compositional Reservoir Simulation of Gasflooding

“…In the real reservoir, there must be residual oil saturation even under the miscible gas injection, S orm , defined by Spence and Watkins (1980). In this paper, the concept of S orm is defined as the residual oil saturation that does not decrease less than physically acceptable values under both miscible and immiscible conditions.…”

“…This residual oil is referred to as miscible flood residual oil saturation (Spence and Watkins 1980). In the conventional compositional simulation, there is no facility to actively define "true" residual oil (nonvaporizing oil) and, hence, the excessive vaporization of oil components into the gas phase is predicted.…”

Unveiling the Effect of Complex Components on CO2-Oil Minimum Miscibility Pressure: Insights from Deep Convolutional Neural Networks

“…The weights W and bias values b are continually updated to minimize the loss. As the pure CO 2 was injected for some experimental measurements of the MMP values, ,,,,,− ,,− ,− certain elements in the 12-element vector pertaining to the contents of specific gas components (e.g., methane, N 2 , H 2 S, and intermediate hydrocarbon) are set to 0. Some optimizers for training the DCNN are not sensitive to minor changes in the gradient caused by sparse vectors, leading to incomplete training of the parameters ( W and b ) and affecting predictive accuracy.…”

“…Data from 225 samples are collected from the available literature ,,,− ,,− (see Supporting Information Table S2). Their statistical characteristics are calculated and shown in Table .…”

Estimation of minimum miscibility pressure during CO2 flooding in hydrocarbon reservoirs using an optimized neural network

“…The precision of model is strongly dependent on adequate data. In this study, 152 experimental values, including 52 pure CO 2 -oil MMPs and 100 impure CO 2 -oil MMPs, were collected from the open literatures (Adekunle and Hoffman, 2016;Bon et al, 2006;Cardenas et al, 1984;Dicharry et al, 1973;Dong et al, 2001;Eakin and Mitch, 1988;Graue and Zana, 1981;Holm and Josendal, 1974;Kanatbayev et al, 2015;Lai et al, 2017;Lashkarbolooki et al, 2017;Li and Luo, 2017;Li et al, 2012;Li et al, 2018;Metcalfe, 1982;Moudi et al, 2009;Harmon and Grigg, 1988;Sebastian et al, 1985;Spence et al, 1980; Thakur et al, 1984;Zuo et al, 1993) to build a robust model for estimating MMP. Eighty percent (122) of the all experimental values are used for training and the remaining 20% (30) are employed for test.…”