A study to assess the value of post-stack seismic amplitude data in forecasting fluid production from a Gulf-of-Mexico reservoir

摘要

This paper describes a study undertaken to appraise the reliability of spatially complex hydrocarbon reservoir models constructed with the use of post-stack seismic amplitude data and well logs. Developments center about the interpretation of data acquired in an active hydrocarbon field in the Gulf of Mexico. The availability of measured time records of fluid production and pressure depletion provides an independent way to quantify the accuracy and reliability of several methods commonly employed to construct static reservoir models. We make use of geostatistical inversion to construct spatial distributions of porosity and permeability that simultaneously honor well logs and post-stack seismic amplitude data. The constructed reservoir models are compared against models constructed with conventional geostatistical procedures that do not make use of seismic amplitude data or else that use a simple statistical correlation between petrophysical properties and seismic-inverted acoustic impedances. We perform multi-phase fluid-flow simulations to assess the consistency of the constructed reservoir models against the measured time record of flow rates of gas/water and shut-in well pressures. For the hydrocarbon field under consideration, the joint stochastic inversion of well logs and post-stack seismic amplitude data yields the closest match to dynamic measurements of fluid production and pressure depletion. We found that the spatial continuity of permeability exhibited the largest influence on the behavior of fluid production and pressure depletion with time. Perturbations to the relationships between porosity, permeability, and acoustic impedance produced significant variations in the simulated time record of fluid production. Finally, we found that the degree of spatial variability of permeability and porosity, and presence of permeability anisotropy were substantially more significant in the prediction of fluid production than variability of relative permeability and capillary pressure.