image

Rezen

Simulation Project Manager

Rezen provides a hierarchical data structure for managing all the simulation runs in a reservoir simulation project. A project contains one or more ensembles, each ensemble representing a scenario the engineer wishes to investigate. Within each ensemble a number of simulation decks can be created, edited and run. The output data vectors can be analysed and plotted on an ensemble or deck basis.

Read More

image

Rezen

Simulation Deck Generator

Rezen provides reservoir simulation deck generation capabilities, using a template based system to process uncertainty or history match variables according to directives inserted in the template file.

Read More

image

Rezen

Simulation Run Management

Rezen provides reservoir simulation run management capabilities, supporting a range of industry standard reservoir simulators.Simulation decks can be submitted individually or as a group. Rezen's own job manager can be tailored to each site or Rezen can be interfaced to your existing queueing system.

Read More

image

Rezen

Experimental Design Algorithms

rezen provides a range of experimental design algorithms that can be used by the engineer to generate new simulation deck input files. Algorithms include Tornado, Plackett-Burman, Central Composite Face, Box-Behnken, Fractional Factorial, Full Factorial, Monte-Carlo etc.

Read More

image

Rezen

Comprehensive Plotting Facilities

Rezen provides a comprehensive range of plot type. For example, any output data vector can be plotted for all the successfully run simulation decks in an ensemble (a collection of related decks).

Read More

image

Rezen

Extensive Plotting Options

Rezen including line style,colour and width, marker style,colour and size

Read More

image

Rezen

Objective Functions

The user can define one or more objective functions to calculate from the output data of  the simulation decks that have been generated and successfully run through the simulator

Read More

image

Rezen

Proxy Simulator, Proxy Decks and S-Curve Plot

Given a user specified objective function, Rezen can build a proxy simulator equation from the output of the experimental design generated reservoir simulation runs. This proxy simulator is a function of the uncertainty variables. By assigning a probability distribution to each uncertainty variable, proxy decks can be generated and objective function values quickly calculated with the proxy simulator. These values are then sorted and plotted as an S-curve from which p10,p50,p90 values can be extracted.

Read More

image

Review

3D Visualization

Review provides 2D and 3D visualization capabilities for viewing and animating reservoir simulation output data

More on 3D Visualization

image

Review

2D Visualization

Review also provides multiple 2D views of the simulation grid and attribute data in either the I, J or K viewing directions.

More on 2D Visualization

image

Review

Local Grid Refinement

Review supports the visualization of simulation grids with grid coarsening and/or local grid refinement including nested LGR.

More on Review Visualization

image

Uncertainty Modelling

Rezen uses experimental design and probability distribution based algorithms to generate multiple reservoir simulation input decks based around a common core input model.

image

Run Manager

Simulation decks can be submitted individually or as a group. Rezen's own job manager can be tailored to each site or Rezen can be interfaced to your existing queueing system.

image

Objective Functions

The user can define one or more objective functions to calculate from the output data of  the simulation decks that have been generated and successfully run through the simulator

What is Rezen?

Rezen is a software tool for managing multiple reservoir simulation runs. Functionality includes :-

  • simulation run management
  • simulation decks generation
  • uncertainty modelling
  • history matching
  • data analysis and post-processing
Rezen aims to support industry standard reservoir simulators, providing a user friendly environment for aiding the design, management and investigation of a reservoir simulation scenario.

A scenario may consist of multiple related instances of a reservoir simulation model, these instances or decks being collectively known as an ensemble.

Rezen supports managing a forward prediction or uncertainty modelling scenario where multiple simulation decks are generated using experimental design algorithms prior to submission to the simulator.

The aim of using an experimental design algorithm is to extract information from a smaller number of simulation runs.

Additionally Rezen supports managing a history matching ensemble where simulation runs are made in serial or parallel, each set of new simulation decks being dependent on the prior runs. The user defines a goodness of fit function as an optimization criterion.

 

Rezen is written primarily in Python and uses the Enthought Tool Suite (code.enthought.com) extensively, in particular, Traits for the GUI, Chaco for line plots, bar charts and 2D plots and Mayavi for 3D. NumPy is used for fast and efficient array handling and SciPy for statistical and optimization functions. 

All these Python packages are open source projects with an extensive number of contributors. Visual Reservoir acknowledges this excellent work. 

Traits can use either Qt or wxPython as an underlying layer. Rezen uses the wxPython layer.

Mayavi uses the VTK visualization toolkit which itself uses OpenGL.

The Python code also interfaces to some Fortran and C legacy code.

A small amount of Tcl scripting is also used. 

A framework diagram is shown below.

image

Page 2 of 2