Dynamic unstructured mesh adaptivity for improved simulation of near­wellbore flow in reservoir ­scale models

Abstract

It is well known that the pressure gradient into a production well increases with decreasing distance to the well and may cause downwards coning of the gaswater interface, or upwards coning of wateroil interface, into oil production wells; it can also cause downwards coning of the water table, or upwards coning of a saline interface, into water abstraction wells. To properly capture the local pressure drawdown into the well, and its effect on coning, requires high grid or mesh resolution in numerical models; moreover, the location of the well must be captured accurately. In conventional simulation models, the user must interact with the model to modify grid resolution around wells of interest, and the well location is approximated on a grid defined early in the modelling process. We report a new approach for improved simulation of nearwellbore flow in reservoirscale models through the use of dynamic unstructured adaptive meshing. The method is novel for two reasons. First, a fully unstructured tetrahedral mesh is used to discretize space, and the spatial location of the well is specified via a line vector. Mesh nodes are placed along the line vector, so the geometry of the mesh conforms to the well trajectory. The well location is therefore accurately captured, and the approach allows complex well trajectories and wells with many laterals to be modelled. Second, the mesh automatically adapts during a simulation to key solution fields of interest such as pressure and/or saturation, placing higher resolution where required to reduce an error metric based on the Hessian of the field. This allows the local pressure drawdown and associated coning to be captured without userdriven modification of the mesh. We demonstrate that the method has wide application in reservoirscale models of oil and gas fields, and regional models of groundwater resources.