Stichting SU2

Mentors: Student: abdulrahman_12

This proposal combines the MLPCpp submodule and Algorithmic Differentiation (CoDi) natively in SU2 for in-solver training, allowing researcher to train their PINNs withing the exact same run instead of training them outside and injecting the weights for inferencing.

Mentors: Student: Thijs Aalbers

The proposed project involves the addition of a pressure-based solver to the SU2 code base. The pressure-based solver is an important algorithm in the Computational Fluid Dynamics (CFD) landscape due to its great performance for low Mach and incompressible flow applications. As of right now, SU2's primary solvers are density-based algorithms which excel in compressible flow simulations. Density-based algorithms can be adapted for low Mach and incompressible flows using for example artificial compressibility or preconditioning. However, these approaches effectively modify a formulation that is not inherently aligned with incompressible flow physics, and therefore remain less efficient than pressure-based algorithms for these applications. The goal of the project is to investigate the current attempts on the addition of pressure-based algorithms to SU2, and to identify what is needed to subsequently work on integrating the pressure-based algorithm into the main SU2 branch. The final result is expected to be a successful merge of a pressure-based algorithm into the main SU2 branch.

Mentors: Student: ddg93

The SU2 code relies heavily on sparse linear algebra, and there is significant speed-up potential in this area through GPU-based processing. In the current implementation, the FGMRES Linear Solver GPU-offloading is being limited by frequent data copies between Host and Device. In this project, I propose to study and implement a truly GPU-native FGMRES linear solver. My idea is to refactor the code so that the largest number of operations is offloaded to the GPU for the duration of the iterative process, all while being compatible with the current level of abstraction and software logic.

Mentors: Student: _Urvi_07

This project improves the SU2 GUI by making it more user-friendly and efficient. It integrates the Python wrapper for smoother simulations, adds real-time visualization, and simplifies solver adjustments. JSON validation helps catch errors early and keeps configurations organized. The installation process is automated to make setup hassle-free, while improvements boost performance and compatibility. Finally, detailed documentation and tutorials ensure users can easily navigate and make the most of the new features.

Mentors: Student: Harsh@2004

This project aims to automate and expand SU2’s testing and validation framework to ensure both numerical correctness and physical fidelity. By sourcing benchmark test cases from public databases (e.g., NASA Turbulence Modeling Resource), we will develop scripts to automatically generate configurations, run simulations, and compare SU2 results against experimental data. Key outputs like lift, drag, and pressure coefficients will be visualized and stored for regression tracking. These validation tests will be integrated into SU2’s CI pipeline, enabling automatic execution, reporting, and performance benchmarking. Additional efforts include modularizing unit tests, improving test coverage, and providing clear documentation for future contributors. This will greatly improve SU2’s reliability, usability, and trust within the CFD community.