High resolution characterization of reservoir heterogeneity with cross-well seismic data

Abstract

Traditionally, reservoir characterization is limited to the inadequate frequency content of surface seismic data and the poor spatial sampling of borehole data for detecting metre-scale heterogeneities affecting fluid flow. Cross-well seismic data can provide the spatial and temporal resolution necessary for imaging these reservoir features. Three synthetic cross-well seismic datasets are created using velocity models that simulate lithologic detail and reservoir heterogeneities at the metre-scale. The first model is derived from an outcrop study of a deltaic depositional environment, and the second and third models are developed from offshore well log data. Statistical analysis of the lateral spatial properties of the high resolution seismic depth images produced from the models provides estimates of the lateral correlation length and the fractal dimension. The results display a unique distribution of spatial properties for each model, indicating that different types of reservoir heterogeneity result in distinctive statistics that are captured by the seismic data. These estimates can be used to provide high resolution constraints on reservoir heterogeneity that can be built into reservoir simulations.