|
Geovisual's 3D visualization software for upgridding and upscaling
The geological model on the left contains five indicator values of which three have been displayed. The model contains approximately 3 million cells. It has been upgridded and upscaled in Rome to produce a grid of approximately 200,000 cells. The directional permeabilities and porosities were read from lookup tables for each of the indicators. The simulation model below displays the K direction permeability. Upscaling methods map the fine grid information stored in a reservoir description model to the coarser scale required by a reservoir simulator such as ECLIPSE or VIP. The aim of rescaling is to provide an effective property distribution over the coarse simulation grid that reproduces the effects of the finer scaled reservoir description grid. Reservoir descriptions are generated on a fine scale to capture the geological heterogeneities as accurately as possible. Reservoir simulators use finite difference methods to numerically model the flow in a reservoir. For computational reasons a reservoir simulation grid is on a much coarser scale than a reservoir description grid. In a reservoir simulation model the finite difference grid is made up of a number of cells which are assigned values for attributes such as porosity, permeability, transmissibility etc. An attribute such as permeability is treated as being homogeneous at the grid block scale. The distribution of permeability over the coarse scale of the reservoir simulation grid must reproduce the effects of the heterogeneities in the reservoir. Rome Features
Rescaling Methods Tensor Based Algorithm for Effective Permeability (KEFF) The Effective Permeability rescaling algorithm is a tensor based algorithm that uses a pressure solver to produce an effective permeability value for the I, J and K directions for each coarse cell. This algorithm uses the coarse cell tangent as the direction of interest and applies a pressure gradient in the direction of the normal to the other two tangents. The averaging of the fine cell velocities is weighted by net rock volume. Porosity/Net-to-gross/Water-saturation The Porosity/Net-to-gross/Water-saturation rescaling algorithm performs a volume-weighted average of the fine cells which lie within a coarse cell. Irregularly shaped corner-point cells are supported. Transmissibility Multiplier The Transmissibility Multiplier rescaling performs an area-weighted averaging of multipliers on the fine cell faces which lie on a coarse cell face. (Note that the fine cell transmissibility multipliers which lie within a coarse cell can be fed into the tensor based effective permeability algorithm in order to influence the effective permeability of a coarse cell). This algorithm uses the same association between fine faces and coarse faces as the KEFF algorithm above. Scalar Rescaling A range of operators may be used on your scalar attributes: arithmetic mean, geometric mean, harmonic mean, variance, mode, number of different values and percentage of mode values. These simple functions give you the power to analyse your data before and after rescaling to give you confidence in the simulation model. Upgridding Rome can construct a simulator grid directly from your geological grid. You control the resolution and Rome has the tools to determine the heterogeneity in your geological grid so you know where the rescaling and upgridding is critical. Data Interfaces Geological Model Input The following data formats are supported for the
input of the fine scale grid and attribute data:
Simulation Grid Input The following reservoir simulation grid formats are supported:
|
  |