Incorporation of fault rock properties into production simulation models
This thesis has two aims. First, to investigate the importance of incorporating the multiphase flow properties of faults into production simulation models. Second, to investigate methodologies to incorporate the multiphase flow properties of faults into production simulation models. Tests using simple simulation models suggest that in some situations it is not particularly important to take into account the multiphase flow properties of faults, whereas in other situations the multiphase properties have proved very important. The differences depend on drive mechanism, well position, and the capillary pressure distribution along the fault as well on the parameters that need to be modelled (e. g. bottom-hole pressures, hydrocarbon production rates, water cuts, etc. ). The results show that it is possible for hydrocarbons to flow across a sealing fault (i. e. 100% water saturation) as a result of its threshold pressure being overcome. The relative permeability of fault rocks may be one of the largest unknowns in simulating fluid in structurally complex petroleum reservoirs. Microstructural and petrophysical measurements are conducted on faults from core within the Pierce Field, North Sea. The results are used to calculate transmissibility multipliers (TMs) required to take into account the effect of faults on fluid flow within the Pierce production simulation model. The fault multiphase flow behaviour is approximated by varying the TMs as a function of height above the free water level. This methodology results in an improved history match of production data. Further, the improved model is then used to plan the optimal time to conduct a follow-up 3D seismic survey to identify unswept compartments. Further, an alternative model was proposed to overcome some of the possible limitations that the previous TM treatments may have at certain stages of a reservoir life. The similar behaviour of the different proposed fault models for the Pierce Field indicate that the current faulting system in this model is not largely responsible for the history mismatch in water production. Multiphase flow properties of faults can be incorporated into production simulation models using dynamic pseudofunctions. In this thesis, different dynamic pseudofunctions are generated by conducting high-resolution fluid flow models at the scale of the reservoir simulation grid block, using flow rates similar to those that are likely to be encountered within petroleum reservoirs. In these high-resolution models, both the fault and reservoir rock are given their own capillary pressure and relative permeability curves. The results of the simulations are used to create pseudocurves that are then incorporated into the up-scaled production simulation model to account for the presence of both the fault and undeformed reservoir. Different flow regimes are used to compare the performance of each pseudoisation method with the conventional, single-phase TM fault representations. The results presented in this thesis show that it is more important to incorporate fault multiphase properties in capillary dominated flow regimes than in those that are viscosity dominated. It should, however, be emphasised that the Brooks-Corey relations used to estimate relative permeability and capillary pressure curves of the fault rock in this study have a significant influence on some of these conclusions. In other words, these conclusions may not be valid if the relative permeability curves of fault rocks are very different to those calculated using the aforementioned relationships. Finally, an integrated workflow is outlined showing how dynamic pseudofunctions can be generated in fault juxtaposition models by taking advantage of the dynamic flux preservation feature in Eclipse 10OTM simulator.