MRST - Transforming research on reservoir simulation

Tutorial lectures
Here we have collected video recordings and slides for a few short courses and tutorials about MRST and reservoir simulation

Øystein S. Klemetsdal: Reservoir modelling and simulation with MRST. Transform 2021, Software Underground, 20 April 2021

webinar-icon.png In the first part of this tutorial, we show you how to get started with MRST, including navigating the modules, constructing a geomodel, and solving simple flow problems. In the second part, we describe the numerical framework, before we advance to more complex problem solving, including multiphase flow simulations and visualization.

 

Knut-Andreas Lie: Building an Open-Source Community Code: the MATLAB Reservoir Simulation Toolbox (MRST). GeoScience & GeoEnergy Webinars, 27 August 2020

webinar-icon.png In this webinar, we review the development of MRST from 2007 and up today:
 # What is MRST and why call it a community code?
 # How and why did it all start?
 # How to make a community code -- going the extra mile
 # Important technological steps on the way:
      - complex grids, procedural libraries and modules,
      - discrete operators and automatic differentiation
      - object-oriented framework (AD-OO)
      - state functions (dependency graph + compute cache)
      - AD backends and high-performing iterative solvers 

 

Knut-Andreas Lie: Introduction to reservoir simulation as practiced in industry, IPAM, UCLA, 2017

This lecture discusses the three basic parts making up a reservoir simulation model: (i) a volumetric grid with cell/face properties that describes the given porous rock formation; (ii) a flow model given as a set of partial differential equations expressing conservation of mass or volumes together with appropriate closure relations; and (iii) a well model that describes the flow in and out of the reservoir.  In particular, we introduce the black-oil equations  and point out some numerical challenges involved in reservoir simulation.

 

Stein Krogstad: Adjoint-based sensitivities in MRSTwith application to multi-segment well completion design, IPAM, UCLA, 2017

  The lecture explains how you can compute adjoint-based sensitivities in MRST. Sensitivities for reseroir controls have been around for a while, but here we also explain how to compute gradient and sensitivities with respect to parameters. We present a simple example of valve model scaling for a multisegment well and an analysis/matching of model parameters for the Sleipner CO2 injection. 

 

Multiscale Methods Summer School, Hasselt University, Belgium, June 2017

As part of a 4-day summer school on 'Upscaling techniques for mathematical models involving multiple scales' organised at Hasselt University, Knut-Andreas Lie gave four talks:

  1. Introduction to flow simulation as used in reservoir engineering
  2. Introduction to the MATLAB Reservoir Simulation Toolbox (MRST)
  3. Flow-based upscaling of reservoir models
  4. Multiscale methods for reservoir simulation

Published November 11, 2017