Rubis-Puma is a 3D, 3-Phase multipurpose, numerical reservoir simulator which extends the KAPPA-Workstation dynamic data workflow to full-field applications.
With an interface intuitively familiar to Rubis users, Rubis-Puma facilitates field-size geomodeller grid import, multiple PVTs and complex multi-region reservoir systems.
Powered by the proven robust numerical engine based on the IFPeN numerical kernel, connection is seamless with other KAPPA modules.
- Defining Geometry
- Reservoir Properties
- Well Properties and Controls
- Dynamic Simulation
Large CPG grids (in formats including GRDECL, CMG, KEG, and KEG5) are imported directly.
Property fields can be loaded separately and assigned to an existing grid.
Geometric fields can be computed from primary fields at load time.
Multiple versions of the same property type can be imported with flexibility to assign which to use for target simulation.
Derived fields can help reduce the impact of data sparsity or serve as a basis for delineating reservoir regions.
It enables computing pertinent fields from other fields based on user-defined rules, such as permeability fields from porosity.
Indicator fields can be derived and used as a proxy to isolate facies and delineate reservoir regions.
Historic data requires QAQC checks and, where necessary, pre-processing before running a simulation. QAQC tools in Rubis-Puma include:
- Data filtering and de-noising
- Data density reduction
- Conversion to bottom-hole pressure
- Pressure and rate synchronization
Black-oil correlations exist to describe fluid behavior.
Given measured PVT data, the models can be tuned to match observation.
Fluid model exports from external PVT packages can also be imported and used directly in simulation.
An equation of state (EOS), including Peng-Robinson and Soave-Redlich-Kwong models, can be used to describe reservoir fluids.
A rich library of pure components and their properties is inbuilt; pseudo-components and their defining properties can be input.
An option exists to preview the resulting fluid envelope. Exports from external packages may be brought into Rubis-Puma.
Using the multi-PVT infrastructure, it is possible to define several fluid models and assign each to a different reservoir trap.
This feature is useful for representing compositional gradients and fluid compartmentalization.
Faults are defined as inter-cell connections which can be imported directly in EclipseTM format or manually created by specifying faces between cells.
The interactive 3D viewer allows modification of fault transmissibility multiplier and fault faces.
Local Grid Refinement (LGR)
LGR options in Rubis-Puma enable focusing grid refinement on the region or location of interest.
It is possible to anchor LGR definition on well trajectories or well perforations and, hence, eliminate the tedium of explicitly specifying grid coordinates.
Imported geomodeller grids can be further streamlined for simulation using robust upscaling options, including automatic, manual, and a hybrid of both.
With indicator property fields in place, creating reservoir regions is seamless: an intuitive wizard guides anchoring region creation on indicator fields or on well trajectories/perforations.
Assigning unique rock and fluid properties to each region is also possible. A 3D viewer and a region editor further enhance the process.
With the provision to load different versions of the same property type, Rubis-Puma also affords the ease of selecting which version of a petrophysical property to use in simulation.
Petrophysical properties can be explicitly defined for matrix and for fractures when modeling naturally-fractured reservoirs.
Traps (Initial State)
The trap property provides the means to specify initial state.
Reservoir initial state may be defined using a single or multiple traps.
Relperm and Pc Data
The governing KrPc model may be imported from a data file in EclipseTM format or created using an interactive editor.
Additional options such as hysteresis, end-point scaling, J-function, analytical Corey coefficients are also available.
Different models are available to compute three-phase relative permeability curves.
Several plot options are available for QC of grid and properties prior to simulation.
Logical operations on grid properties may be carried out using, for example, the fence diagram and filters.
A statistics viewer is available to visualize statistical distributions of imported and created grid properties.
Well Data Load
Well trajectories, perforations and logs can be imported directly from standard data files (including ASCII, LAS, LIS, DLIS, etc.).
Production/injection history can be imported from standard data sources for all wells at once using the multi-well data import option.
It is possible to manage and/or QC imported pressure and production data via the interactive history editor as well as perform wellhead-bottomhole pressure computations.
Either for operational or reporting purposes, wells can be easily grouped together interactively.
This enables treating them as a single entity in model set-up as well as in visualizing results plots.
A multi-level organization exists for setting up production constraints on a well, group or field basis.
Additional constraints can be defined for well abandonment conditions.
Simulation can be run on remote agents on both Windows and Linux platforms.
Various numerical solver algorithms optimized for large and super-large grids are available.
The restart capability provides a means to initiate simulation from a stored intermediate state, thus saving time and enabling the user focus on incorporating new operating conditions.
Several tools are available for visualizing simulation results, including graph plots, 3D views, and animation.
Results can be visualized at well, group and field levels.
The data store enables creation of static copies of data or simulation results for comparison or reporting purposes.
Results may also be exported (as raw data or graphics) for further processing and/or use in external packages.
No limits exist on the number of simulation cases possible within Rubis-Puma except those imposed by the hardware.
To reduce this impact, the memory management infrastructure provides the means with which to load/unload simulation cases from memory.
This helps to keep the demands on computational resource to a minimum and thereby enhance overall performance.