The Study Of Modeling Constrained By Seismic Data And Facies

Conventional methods cannot quantitatively describe the facies of fluvial petroleum reservoir due to its small-scale structure and unstable lateral variation and poor reservoir heterogeneity. At present, the main method to characterize the reservoir heterogeneity is to build the model based on geostatistics. The research hot spot is facies controlled modeling. The conventional method first uses log data to build the facies model so-called facies modeling. Then the physical property in model led based on the facies. This is facies controlled modeling. However, the first step can be improved due to that the facies model derived from facies modeling has uncertainty. In general, deterministic facies model through geological study can be available. In this study, we use the deterministic facies model as constraints to study the property modeling methods, such as Co-Kriging and Sequential Gaussian Simulation which can integrate the seismic and log data. In order to validate the method, a synthetic model is built. The study starts with variogram study based on spatial distribution of the model data before and after constrained by facies. Then Co-Kriging and SGSIM are used to build the model. Models built using these two methods constrained by facies are compared with the realizations derived from methods without constraints. The comparison indicates that the realizations with improved facies constraints follow the geological model and accuracy is improved. In addition, the modeling method is applied to a field case for further validation.