A Novel Method for Multi-Resolution Earth Model Gridding

Abstract

Abstract Earth models are important tools for support of decision making processes for optimal exploitation of subsurface resources. In earth modelling applications, massive amounts of information result from collecting and interpreting measurements and simulating physical scenarios at various scales and resolutions. The size of an earth model grid is a consequence of its scale and resolution. The grid size is a major factor for the computational efficiency when managing the model and performing various model based simulations. The grid is controlled by the geological structure, and the resolution of the grid is therefore controlled by the resolution of the geological structure (the number of structural elements). We discuss principles for a multi-resolution framework for earth model gridding where the aim is to allow automatic, local control of the resolution of a populated grid by locally controlling the resolution of the geological structure. The structure is ordered in a hierarchical manner, and splits the subsurface into a nested set of regions that are ordered in a corresponding hierarchical fashion. Each region can be separately gridded. The method is based on a recent method for local updates of the connectivity (topology) of the geological structure and of the resolution of a populated grid. Similar to existing strategies for multi-resolution modelling in other sciences and applications, the proposed method allows local control with the trade-off between numerical accuracy and grid size. The method should allow to locally change the model resolution even during a modelling exercise; subsurface volumes of high interest can be represented at fine resolution, whereas in volumes of less interest, details can be omitted. The method also enables local updates and local scale uncertainty management of the geological structure. The aim of the proposed method is to develop an effective methodology that supports real-time earth model based workflows such as geosteering where multiple grid realizations are never fixed and always updated with the most recent measurements and interpretations, and where each realization is always kept at an optimal resolution.