Reservoir characterization analysis in glacial reservoirs

This research aims to characterize reservoir properties by applying rock physics and AVO analysis followed by pre-stack inversion. Two approaches are investigated: One approach addresses the case in which there are wells and seismic data, and the other addresses cases where only seismic data are available. The former approach is achieved by using well-log cross-plots for rock physics modeling to determine the feasibility and pay zone through gas fluid substitution followed by AVO analysis. Pre-stack inversion is then used to predict porosity and gas saturation. In the second approach, a synthetic seismogram is generated and compared to the observed seismic trace at the location of interest by forward modeling P-wave interval velocity and density. The best-matching P-wave velocity and density are subsequently used to generate synthetic well logs at the same location. Pre-stack inversion is then performed on these synthetic wells to predict porosity and gas saturation. Property prediction is performed by a feasibility study and pay zone calculation using rock physics modeling of the nearest well to the seismic block. Finally, the results of this case are validated using real wells. This new approach of reservoir characterization using synthetic wells is applied on reservoir channels and yielded a fairly good porosity prediction but a less accurate prediction of gas saturation.

Automated summary

This research focuses on characterizing reservoir properties using rock physics and AVO analysis, followed by pre-stack inversion. Two approaches are investigated: one utilizing well and seismic data, and the other utilizing only seismic data. The first approach involves well-log cross-plots and AVO analysis to determine feasibility and pay zone, followed by pre-stack inversion for predicting porosity and gas saturation. The second approach involves generating synthetic seismograms and synthetic well logs for predicting reservoir properties. The results are validated using real well data.