| Proper modeling of water injection projects in undeveloped deepwater reservoirs is important not only from a recovery factor point of view, but also for environmental protection purposes. Around the world some wells producing from reservoirs under water injection were shut down because of oil leakage to sea. Oil reservoirs with major faults that communicate the reservoir to the seabed are sensitive to water injection projects. Oil companies' common practice is to reuse the produced water and to reinject it into the reservoir. This procedure may cause pressure increase, and also may result in fault slip and communication through the fault from the reservoir to the seabed. Therefore, the flow simulation must consider concepts of shear and effective normal stress to estimate the maximum allowed pressure that generate slip in the fault, and, consequently, leakage to the seafloor. The main objective of this research is to explore the production uncertainties in an undeveloped oil reservoir, under water injection project, with a major internal fault, and the role of geomechanics in the fault reactivation.;This research developed a new external iterative computational program applied to complex geological geometry that updates geomechanical information for each cell grid, for each time-step and during each Newton's iteration, to estimate the maximum pressure allowed to safely operate a water injection project without any oil leakage to the sea. This new methodology links two commercial simulators: one that emphasizes flow through porous media aspects, and the other that focuses on the geomechanical behavior. For this particular field, the maximum allowed water injection pressure was estimated in a value similar to the initial reservoir pressure. Sensitivity analyses of reservoir temperature, injection pressure and well location on the fault were carried out, and a comprehensive stress map and oil leakage variation, associated with flow through the fault, was also obtained. The investigations were limited to numerical simulations, but they used seven triaxial laboratory tests performed in Brazil as support information. |
