Developing a workflow for a study of polymer flooding in heterogeneous reservoirs

Polymer flooding can significantly improve the sweep efficiency in heterogeneous reservoirs compared to a water flood. This is caused by the improved mobility of the polymer flood, which has an effect on the crossflow mechanisms caused by viscous, capillary and gravitational forces in the reservoir. However it is not yet fully understood how the different physical and geological parameters which play a role in polymer flooding in heterogeneous reservoirs, interact with each other. Therefore a workflow has been developed to perform a sensitivity study on the influence of different parameters on the performance of a polymer flood. This workflow incorporates the ability to optimize well control parameters based on the project Net Present Value. The workflow is applied to perform a sensitivity study on two different types of models in order to study different types of heterogeneity: 2D layered reservoir models and 3D simplified turbidite reservoir models. The parameters of the sensitivity study are based on a literature review, from which research gaps are defined. The 2D study is performed with 1300 different models with varying physical parameters and architectures with the same permeability distribution. The following results have been found: Firstly, the effect of different oil viscosities is influenced by the optimized flowrate. As a result the pressure constraints have a larger effect on the models with a high oil viscosity than on models with a low oil viscosity. Consequently the models with a high oil viscosity have less benefit from a polymer flood. Secondly, the capillary pressure has a significant effect on the performance of a polymer flood compared to a water flood, despite of assumptions made in literature that capillary pressure may be ignored. The most important influencing parameter for this 2D study is the geology of the model. This study shows that heterogeneity factors such as the Dykstra-Parsons factor are not sufficient to capture the effect of geology on the performance of both a polymer and a water flood. Hence a new heterogeneity factor has been developed, which captures the trends visible from this study fairly well: the Sequence Factor. This factor incorporates differences in layer ordering, layer thicknesses and vertical permeability ratios. The 3D study is based on 9 different realizations of a simplified channelized turbidite reservoir model, consisting of channels and surrounding levees with different properties. The realizations are obtained by varying geometrical input parameters. Two different optimization runs have been performed on one of these models: an optimization which includes optimizing the flowrate and an optimization in which the flowrate is not optimized. Instead the flowrate is fixed at a maximum of 1.5 pore volumes of water injection over 10 years. The other 8 models are simulated with the optimized input parameters of the latter optimization. The results of this 3D study show that when the flowrate is optimized, the polymer flood is affected by the pressure constraints. Hence the polymer flood has almost no benefit compared to the water flood. When the flowrate of the simulation is fixed, the benefit of a polymer flood is significant. For that model the pressure constraint is not impacting the polymer flood, which leads to the flowrate staying constant instead of decreasing. Simulation of the other 8 models show that 1) well placement in a non-optimal location with regards to connectivity to the reservoir may lead to injectivity issues during the polymer flood, and 2) that the connectivity in the geological model itself may have an influence on the performance of the polymer flood. In order to find a proper relation between reservoir connectivity and the performance of a polymer flood, more data are required.Petroleum EngineeringGeoscience & EngineeringCivil Engineering and Geoscience