Experimental, numerical and economical analysis of polymer floods in stratified reservoirs

Abstract

Polymer flooding, whether in secondary or tertiary mode, is one of many enhanced oil recovery (EOR) processes that will be more heavily relied on in the future to extract oil more efficiently from reservoirs. Polymer flooding reduces the mobility ratio and therefore accelerates the production of oil from reservoirs by improving the fractional flow which consequently reduces channelling, bypassed oil and water cuts. In communicating stratified reservoirs, polymer flooding can improve vertical sweep in addition to areal sweep. Vertical sweep in stratified reservoirs is improved when polymer is injected as this alters the pressure profile in each layer. As a result, viscous crossflow is induced whereby oil from the high permeability layers crossflows to the low permeability layer which therefore improves flood front conformance and delays water breakthrough. In this thesis, we to first derive a set of 10 dimensionless numbers that describe polymer flooding in stratified reservoirs using inspectional analysis. The analysis considered both capillary pressure and polymer adsorption. Four new dimensionless numbers not reported in the literature were produced, 1) a modified mobility ratio 2) dimensionless capillary pressure numbers 3) a modified capillary number 4) adsorption number. All dimensionless numbers produced through inspectional analysis were validated using numerical simulations. Second, we validate ECLIPSE 100’s polymer flooding model by comparing its predictions with the results from a set of carefully designed series of laboratory displacement experiments using glass beadpacks. Water and two glycerol solutions with different viscosities were used as displacing fluids to displace paraffin from the glass beadpacks. Paraffin and glycerol solutions serve as analogue fluids to viscous oils and polymer solutions respectively. Displacement experiments were initially done in homogenous packs to measure the static and dynamic petrophysical properties of the glass beadpack such as porosity, permeability and relative permeability. These displacement experiments were performed using two bead sizes. The two bead sizes were then used to construct the two-layered glass beadpacks. In one beadpack, complete vertical communication was allowed whilst in the other pack the communication between layers was prevented by placing a rubber barrier placed between the two layers. In separate experiments, using water and glycerol solution, paraffin recovery curves were obtained for both the communicating and non-communicating beadpacks. The recovery curves from both beadpacks allowed us to measure the incremental paraffin produced due to crossflow and mobility control after 1 PVI of either water or glycerol solution. The ECLIPSE 100 simulation model included all the properties measured in the lab and the output was compared to laboratory results. The comparison between laboratory and numerical simulation showed good agreement and thus validated the polymer flooding model. The good agreement between laboratory and numerical simulations gave us the confidence to move forward and investigate wider ranges of mobility and permeability ratios on the lab scale. The linear displacement in the laboratory scale simulation models indicated that maximum incremental oil due to crossflow occurs at moderate permeability ratios and favourable mobility ratios. Finally, the impact of permeability and mobility ratio on the incremental oil recovered due to crossflow was investigated on the field scale. A ¼ five spot pattern was used with both idealised and realistic polymer slugs. The results echoed the trends of incremental oil due to crossflow seen on the laboratory scale, however, the magnitude of crossflow seen on the field scale was less than that seen in linear displacements on both the laboratory and field scales. This was due to the pressure distribution in the reservoirs for each displacement type. In addition, idealised polymer slugs (as expected) showed higher incremental oil due to crossflow when compared to realistic polymer slugs. This is because the idealised slugs maintain their viscosity as they travel from the injector to the producer unlike realistic slugs. However, realistic slugs still produce significant amounts of incremental oil due to crossflow. Increased model heterogeneity was investigated by using a sector model extracted from the SPE10 model 2. The results showed that as heterogeneity of the reservoir increases so does the incremental oil due to crossflow. The financial impact of this incremental oil due to crossflow was evaluated by using time value of money and calculating the present value at different mobility and permeability ratios.