Juliacon 2024

Variational Multiscale Method in Julia for airfoils simulations
2024-07-10 , While Loop (4.2)

A stabilized finite element method - the Variational Multiscale Method - has been implemented in Julia. It is based on Gridap library, and the parallelization on GridapDistributed and PartitionedArrays. It has been tested to simulate airfoil up to Reynolds 500 000.


The Variational Multiscale Method has been implemented in Julia relying on the Gridap ecosystem. It solves the Navier Stokes equations using the finite element method with same-order interpolation elements. Thanks to Gridap we can solve the non-linear, time-dependent problem on distributed architecture with satisfying strong and weak scaling. To speed up the computation also a linearized version is implemented. The final system is solved using PETSc. Thanks to Julia it is possible to write equations almost the same as on paper, differently from other programming languages.

The main application is to perform implicit LES on transitional airfoils (Reynolds up to 500 000). It has been validated on the sd7003 airfoil and experimental results.

It can be shown how easy is also to implement the continuous unsteady adjoint since the PDEs are available.

repos:
https://github.com/carlodev/SegregatedVMSSolver.jl
https://github.com/carlodev/IncompressibleAdjoint.jl

Videos:
https://www.youtube.com/watch?v=wyJnVTuCC3Q
https://www.youtube.com/watch?v=manQpEMwj4Y

See also: GitHub