A new pore-scale numerical simulator for investigating special core analysis data
Boujelben, Ahmed Hamdi
MetadataShow full item record
The study presented in this thesis addresses various pore-scale phenomena related to the oil industry by implementing new numerical models capable of simulating a wide range of multiphase flow processes such as depressurisation, water flooding, gas injection and various EOR techniques. The aim is not to produce quantitative predictions per se but rather to examine the effect of key petrophysical parameters on oil recovery when different production protocols are applied to specific rock analogues. In order to facilitate this, a new pore-scale process simulator is developed – numSCAL (numerical Special Core Analysis Laboratory) – with different modules associated with different mechanisms. A steady-state depletion model is described first and used to investigate the impact of numerous parameters on solution gas drive. We show that parameter combinations that increase bubble density can lead to delayed gas breakthrough and can result in high critical gas saturations. The model is extended to support three-phase flow by incorporating concepts from graph theory. Simulation results highlight the interaction between the underlying phase saturations, spreading conditions and wetting films and emphasise the competition among mechanisms acting in three-phase systems. Two unsteady-state models are also presented to study water flooding processes in porous media – the first mainly applied to simulate drainage processes and the second used to study the onset of ganglia mobilisation. Results show that parameters affecting the capillary number and viscous ratio play a crucial role in determining the observed invasion regime and final oil recoveries. Conditions required for ganglia mobilisation are derived and used to predict the likelihood of mobilisation at different parts of the reservoir. The dynamic drainage model is then extended to simulate low salinity (LS) water flooding and polymer injection – secondary and tertiary effects are shown to depend on interactions amongst several key flow parameters (including initial reservoir wettability, flow rate and viscous ratio). In addition, a positive synergistic effect is identified, where the combined injection of LS brine and polymer leads to increased recovery in several scenarios. The study concludes with an application of the pore-scale modelling technique in a novel research area. A new approach is presented to model drug perfusion surrounding Glioblastoma Multiform (GBM) tumours. Results show that blood flow, transmural transport and tissue diffusion have a direct impact on the average drug concentrations that develop in the vascular network and the surrounding tissue.