One key aspect in the management, planning and analysis of the environmental impact of oilfield exploitation is the reliable prediction of behaviour. Computational models have two basic elements. The first is advanced algorithms to generate high order meshes that accurately represent the complex geometry of the oilfield (definition of boundaries, localisation of faults, cavities and strata, injection and extraction wells, etc.). The second is robust, efficient numerical methods to predict the behaviour of the oilfield highly accurately in terms of time and space.
The new numerical techniques developed by LaCàN resolve the difficulties of the physical models that are currently used that can lead to inconsistencies or considerable precision errors. For example, near extraction wells and around faults, there are large gradients of pressure which give rise to larger changes in velocity. In order to correctly capture these abrupt variations, which may significantly affect the reliability of results, commercial software tends to generate excessively fine meshes, because they rely on first order methods. This leads to (often prohibitively) high computational costs to obtain predictions with the required precision.
The high-order hybridisable discontinuous Galerkin method (HDG) is designed to overcome these issues. First, the new numerical technique improves the discrete representation of the oilfield through a new algorithm for generating high order 3D meshes of high quality to discretize the geometric model of an oilfield. Second, the physical model is resolved through a formulation that combines a high-order hybridisable discontinuous Galerkin (HDG) method with schema of temporal integration that are also high order. This combination enables an increase in precision of flow simulations for an oilfield.
The project is supported by the Catalan Institute for Energy Research (IREC) and PetroSoft, as observing promoting entities (EPO), and has received funding from the Spanish Ministry of Economy and Competitiveness.