ABSTRACT: Seismic history matching and the influence of seismic reprocessing on reservoir performance prediction
Bayly, Martin1, Xuri Huang2,
and Geoff King3
(1) Western Geophysical, Perth, East Perth, WA, Australia
(2) Western Geophysical, Houston, TX
(3) Western Geophysical, N/A
Once an oil reservoir is being produced, it is important to understand the fluid movement, the flow mechanism, and the reservoir heterogeneity. Recently 3-D time lapse seismic data have been successfully used to improve reservoir characterization and management. These mostly qualitative studies showed that time lapse seismic even on legacy data resulted in a better understanding of the production behaviour of the reservoir (Anderson et al. 1997, Eastwood et al. 1998). Huang et al. presented a quantitative seismic history matching method which integrated time lapse seismic data with reservoir simulation (Huang et al. 1998a). The approach uses a stochastic optimization method to minimize the differences between synthetic and observed time lapse 3D seismic data and simultaneously matches the production history at the wells. Huang et al. (1998b) further improved the seismic history matching methodology by deriving a more representative initial porosity estimation and by optimizing the relative permeability curve.
Using two 3D seismic data sets acquired over a field in the Gulf of Mexico, this paper reviews the seismic history matching method and demonstrates how data reprocessing will affect the reservoir modeling, especially the quality of the dynamic performance prediction. Based on 'off-the-shelf' data and simple post stack matching a reasonable difference map is obtained. An initial reservoir model is characterized using the base survey and the model is updated using the seismic history matching method. Dynamic performance in the time frame after the monitor survey, although more accurate than that based on conventional history matching, underestimates actual production. It is then demonstrated how reservoir performance prediction is much improved through the use of pre-stack seismic reprocessing of the two 3-D volumes before computation of the seismic differences.
AAPG Search and Discovery Article #90913©2000 AAPG International Conference and Exhibition, Bali, Indonesia