Growthism: A Complete Framework for Integrating Static and Dynamic Data Into Reservoir Models
Access full-text files
Accurate characterization of subsurface oil reservoirs is an essential prerequisite to the design and implementation of enhanced oil recovery (EOR) scenarios. Specifically, in reservoir characterization, integrating static and dynamic data into reservoir models to construct accurate and realistic models has received considerable attention. Unlike most of the conventional geostatistical approaches of integrating data into reservoir models that are based on semi-variograms (two-point statistics) as a measure of spatial connectivity, a complete multiple-point statistics framework is presented in this dissertation. In contrast to two-point statistical methods, multiple pointstatistics-based methods are capable of reproducing curvilinear geological structures. The algorithm starts with extracting multiple-point (mp) statistics from training images (conceptual geological description) using an optimal spatial template. After collecting different patterns (data configurations) and building the multiple -point histogram, the pattern reproduction process commences. It begins from data locations (simulatable nodes) and then grows to fill the entire reservoir domain. The algorithm accounts for three main practical issues: uncertainty in geological scenarios, deriving scanning template and non-stationarity. The current algorithm, unlike others, for the cases with many possible geological scenarios ranks the training images based on the consistency between the training images and hard conditioning data. A fast and robust algorithm to derive optimal spatial templates is presented that is based on a semi-automated procedure. Results prove that pattern reproduction using the optimal template is better than using just an arbitrary template specified by the user. Growthsim is capable of integrating data from multiple data sources. Non-stationarity, in terms of variations in facies proportion, can be represented and synthetic and field examples are presented in this dissertation. The conventional approach to integrate production information into reservoir models is by iterative perturbation of the reservoir model until the production history of the reservoir is matched. Iterative methods have been applied to random fields that are completely characterized by a two-point covariance function. An alternate novel technique, as implemented in this research, is based on the merging of multiple-point statistics inferred from history matched and geological models. Pattern growth is performed subsequently by sampling from the merged multiple-point histograms. History matched models, using the presented approach, show an excellent agreement with underlying geological descriptions and match production history. It is demonstrated that the procedure yields a reliable reservoir model best suited for flow prediction.