Inverse Reservoir Modeling Using 3-D Seismic

Abstract

A new software system is proposed with the objective of obtaining detailed reservoir delineating and characterization on the basis of deviated well log data, seismic time interpretation, geologic constraints, and three-dimensional (3-D) stacked seismic data. Inverse modeling is used as the technology to achieve this. Advantages of inversion are the integration of all available information and extensive quality control (e.g., resolution and variance analysis). An automated two-step method is proposed. First, the parametric model is tied to the wells and the seismic interpretation. Second, inversion is used to fine-tune the parametric model between the wells using the 3-D seismic data. The underlying parametric model should be as sparse as possible to guarantee a stable and efficient inversion. The detailed reservoir characterization is described by a grid cell model that is built from a stratigraphic framework and a set of typical pseudo wells per stratigraphic sequence. The stratigraphic framework includes erosional surface, onlaps, and pinches. The set of typical pseudo wells is found by analyzing the actual well logs using Singular Value Decomposition (SVD). SVD is a transformation that uses all the statistical information (variance, mean and dependencies) present in the data. The net effect is a parametric model that implicitlymore » handles all relations between log data (e.g., Gardner) without making any assumptions. Seismic relations (such as convolution and multiples) can be handled by including seismic data in the SVD.« less