Given a well test, the aim of this paper is to demonstrate that the range of possible reservoir boundaries derived from a seismic image is a valid constraint in the reservoir response.
A lens model was considered. Synthetic seismic and well test data were generated. The focus was to verify that the single phase well test response is sensitive to modifications of the reservoir boundaries derived from a poststack seismic depth image, all the other variables being fixed.
A specific prototype software was developed to implement the proposed methodology.
Reservoir geology, geophysics and engineering (GGE) data each contain different measures of reservoir properties. Traditionally, the most formal way of estimating these properties is to invert each data separately to recover partial information of the reservoir. Such results generally suffer from strong non-uniqueness which is overcome by introducing often non-trivial prior information. Typically, interpretation of logs and laboratory measurements on cores and plugs gives detained 'hard' information of lithofacies and petrophysics at the wells.
For instance, geological interpretation in terms of high resolution stratigraphy gives information on the lateral extension of the lithofacies. On the other hand, 3D seismic imaging provides continuous indirect and averaged information of the reservoir in terms of acoustic/elastic parameters. Finally, reservoir engineering data provides understanding of the fluid flow characteristics.
Any viable integration of GGE data relies on the trivial but basic assumption that reservoir information gained by combining these data is greater than that derived by using a single type of data. The range of benefits would be to provide at best a consistent GGE reservoir image reducing non-uniqueness, and at least a non-conflicting representation of it. Within this framework, one could state schematically the contribution of each type of reservoir information: geology would provide the 'static interior' of the reservoir at finite locations, geophysics would yield a blurred spatial image of the reservoir such as a 'fuzzy skeleton' (e.g., shape, extension) and possibly some petrophysical constraints, and finally, engineering would characterise flow parameters of the reservoir.
Scarce literature exists in the area of integration of seismic data with fluid flow simulations. On 3D synthetic poststack seismic data, Araktinghi and Bashore have investigated the effects on flow behaviour of varying seismic parameters such as frequency bandwidth, waveform phase estimation and correlation level between seismic and well-log data.