Jutul
Experimental Julia framework for fully differentiable multiphysics simulators based on implicit finite-volume methods with automatic differentiation.
Numerical solvers and solution algorithms
Numerical solvers and solution algorithms are central to simulation, enabling the solution of large, coupled systems of equations that arise from physical models. Research in this area ensures robust, scalable, and efficient simulations for complex, nonlinear, and time-dependent problems.
Contact persons
Proficiency in nonlinear solvers and solution strategies is broadly applicable across simulation domains. However, the formulation and implementation of robust and efficient methods vary significantly depending on the physics, coupling, and time scales involved. Developing new and effective algorithms therefore requires deep insight into the mathematical structure of the models and the computational challenges specific to each application.
What do we do?
We develop and implement a wide range of solver technologies, with a focus on methods that are robust, scalable, and tailored to the needs of multiphysics simulation. Our expertise centers around the following topics:
- Iterative solvers for large sparse systems, including Krylov subspace methods, multigrid techniques, multiscale methods, etc. We also work on GPU-acceleration.
- Preconditioning strategies for coupled and multiphysics problems, including block-structured and physics-based approaches.
- Nonlinear solvers, such as Newton-type methods, trust-region strategies, and methods for efficient localization of computations for nonlinearly stiff problems including nonlinear domain decomposition and optimal ordering.
- Solution strategies for coupled systems, including fully implicit, sequential, and multilevel methods.
- Adaptive time-stepping and solver control for stiff and time-dependent problems.
- Parallelization using distributed and shared memory paradigms as well as GPU acceleration to ensure scalability on modern hardware.
Software
MRST - MATLAB Reservoir Simulation Toolbox
A free open-source community code for rapid prototyping of new methods for modelling and simulation of flow in porous media. Has a large user community from all over the world.
Open Porous Media (OPM)
The Open Porous Media (OPM) initiative provides open-source software for simulation, upscaling and visualization of porous media processes, in particular subsurface reservoirs.
Selected projects
Improved methods for reservoir simulation
Reservoir simulation, an essential tool in the oil and gas industry for predicting reservoir behavior, often incurs significant computational expenses. This arises from the intricate nature of subsurface flow dynamics and the need for high-fidelity...
Tools for model-reduction and accelerated reservoir simulation
To optimize hydrocarbon reservoir recovery, understanding and predicting flow and transport processes is crucial. Geo-cellular models, representing complex rock formations, often contain millions of cells, requiring hours for simulation. To expedite...
Multiscale performance
Multiscale methods, such as MsRSB initially proposed by SINTEF, have emerged to enhance runtime efficiency and pressure solve scaling in reservoir simulators using sequential splitting. Starting from the methods successfully implemented in the...
Full-field simulation of water-based EOR
We study and develop numerical tools that can be used to improve the resolution of EOR simulations and, in particular, capture accurately the impacts of the injected chemicals on the recovery process.
Next generation multiscale methods for reservoir simulation
This project strives to broaden the scope of multiscale technology by delivering substantial acceleration to conventional reservoir engineering workflows. By doing so, it not only facilitates a considerable speedup in traditional processes but also...
Multiscale-streamline simulation of highly heterogeneous and fractured reservoirs
The purpose of the project is to develop a prototype multiscale mixed finite-element pressure solver in the ECLIPSE FrontSim streamline reservoir simulation software. FrontSim is a three-phase, 3D simulator that models multiphase flow of fluids along...