Multiscale Flow Simulation Research Articles

Taking into account the complexity of natural fracture systems in reservoir single-phase flow modelling

In order to develop a methodology for reservoir flow modelling that takes into account the complexity of natural fracture systems, high-quality fracture datasets were collected at different scales in a well-exposed clastic formation at Tayma, Saudi Arabia, and used to design a multiscale fracture model incorporating both large-scale faults and medium- to small-scale fractures (joints) for numerical simulation. Flow modelling was then conducted, taking into account three scales of permeability to represent the matrix, joints, and faults. Actually, the most severe situation was considered, where the permeability of the joints dominates strongly the one of the matrix. The results obtained with this model show that, on a local scale, the maximum directional flow is always horizontal along the direction of a systematic joint set—in this case NW–SE. Vertical flow is only possible where joints j and bedding planes b have conductivity ratios C j/ C b between 0.1 and 10. The flow properties of the model were then upscaled to simulate the entire reservoir through combining the flow properties of the jointed matrix (represented by an equivalent permeability tensor) with those of the large-scale fault system (which was modelled explicitly) in a dual-permeability system. Simulation of a pumping well to illustrate the hydraulic response of the multiscale fracture system according to different permeability ratios between the faults and the jointed rock matrix, i.e. K f/ K m, showed that the influence of the fault system begins to be visible at a permeability ratio of 100. At a ratio of 10,000, the hydraulic potential field is controlled completely by the fault system, the fault planes becoming isopotential lines of the hydraulic field. This multiscale fracture flow simulation procedure, which combines the hydraulic effects of small- to medium-scale fractures and large-scale faults, could be very helpful in predicting flow responses in dual-permeability reservoirs.