Abstract

In petroleum exploration, seismic amplitude information is used to delineate the lithology type and the presence of fluid. To infer the detailed rock and fluid properties, well data has to be carefully and closely integrated with the seismic data. A successful reservoir characterization mainly includes derivation of rock properties from the seismic data and relates them to reservoir properties and mapping their distribution within the reservoir. Post-stack seismic inversion has been carried out to understand various rock properties from the seismic section. To delineate and resolve sand reservoirs accurately, an approach has been made here to extract the reservoir properties utilizing Multilayered Feed Forward Neural Network (MLFN) technique. This technique is applied in geologically challenging Upper Assam basin of North-East India for estimating volume of shale (Vsh). By knowing the Vsh distribution, other reservoir properties can be properly estimated in the reservoir. The method is proven successfully in this study area with better correlation results between the actual and predicted