Finite-Element/Rapid-Response Approach Promotes Collaboration

Abstract

Listen

Translate article

Technological advancement, particularly in exploration and development efforts, has often left unfulfilled the promise of geoscience and engineering working together more collaboratively and efficiently. But a new approach that combines finite-element (FE) technology with a “rapid response” approach to reservoir modeling is encouraging geoscientists and engineers to shift workflow dynamics from autonomy to collaboration. The new technology, which employs true unstructured 3D mesh, allows modeling time to be reduced from months to weeks, broadening the application to reservoirs once considered impractical. Rapid-response reservoir modeling (similar to top down) starts with the simplest model to obtain all pertinent reservoir information: boundaries, volume, skin, drive mechanism, porosity, permeability, flow barriers, etc. These models are built by employing Resolve, a software package that synchronizes all traditional analyses: material balance, trend analysis, pressure-transient analysis, and simulation. The tool honors full fluid-flow physics, achieves extreme precision when necessary, and iterates rapidly to allow testing of all reasonable scenarios. Houston-based Object Reservoir has used the technology in association with more than 50 companies. Conventional reservoir simulators, originally developed in the 1960s, use finite-difference (FD) approaches for solving fluid-flow equations. FE methods have been developing for nearly 65 years and already have been applied in diverse industries including automotive, aerospace, manufacturing, and construction. Until recently, FE methods were not applied in the oil and gas industry. Some early attempts failed because of a lack of theoretical foundation for handling the nature of fluid flow. In the late 1970s, pioneering work by a variety of researchers led to FE formulations that were stable and accurate for fluid flow. FE methods are easily contrasted with FD, which can be thought of as a collection of gridblocks with distinct properties and step changes at each boundary. The FE method uses spatial integration mathematics with hexagonal or tetrahedral elements. Over each element, a linear variation of the pressure and saturation solution is assumed. This produces a smooth transition of reservoir properties within elements and across boundaries. Then, the equations for governing the flow of oil, gas, and water hold on average over each element, thus conserving mass. The result is a system of equations for values of pressure and saturation at the vertices of the elements (as opposed to gridblock averages in FD methods) that can be solved using standard well-established procedures. FE methods allow for a richer and more mathematically rigorous representation of flow characteristics. FE methods enable advanced computational capabilities naturally. The unstructured grids are useful for boundary-conforming and adaptive meshing, elimination of grid-oriented effects, and advanced visualization. The oil and gas industry is beginning to embrace this technology because of its flexibility, accuracy, ease of use, and workflow/productivity efficiency. Equally as important as the technology is the process employed to use it. The Dynamic Reservoir Characterization process is collaborative in nature. It facilitates communication and interaction between geoscientists and engineers through all phases of the asset life cycle. Enabled by the technology, the process reconciles reservoir-characterization differences between geoscientists and engineers and unites asset teams around all of their knowledge, yielding a common decision-making platform. Traditional modeling workflows are often best characterized by one-way data transfers from the geologist/geophysicist side to the engineering side. However, a highly collaborative, multidisciplinary process uses a “continuous improvement loop” to converge on the solution.