Solver
A
c
c
u
r
a
c
y
S
c
a
l
a
b
i
l
i
t
y
P
h
y
s
i
c
s
I
/O
R
o
b
u
s
t
n
e
s
s
E
f
f
i
c
i
e
n
c
y
F L E X I B I L I T Y
ASPIRE Flexibility is the raison d’être of UCNS3D. A numerical framework that approximates partial-differential equations with high-order Accuracy. The solver is benchmarked on Europe’s most powerful high-performance computing facilities, demonstrating great Scalability up to hundred thousands of computer cores. UCNS3D solves the governing equations of Computational Fluid Dynamics, the Physics of turbulence is captured with state-of-the-art models. The solver can read and write based on the most popular I/O data formats. The Robustness of the solver lies in its foundation, the Finite Volume k-exact framework. UCNS3D went through several optimisation iterations to ensure competitive computational Efficiency. Finally, the solver is formulated for hybrid unstructured grids, this ensures the necessary FLEXIBILITY of industrial scale applications.
UCNS3D Features
Mesh
- Conforming 2D and 3D unstructured meshes consisting of triangles, quadrilaterals, tetrahedrals, hexahedrals, prisms and pyramids.
Solvers
- 2D/3D Linear Advection.
- 2D/3D Compressible Euler, RANS, Laminar Navier-Stokes.
- 3D Compressible DES, DDES, ILES.
- Numerous variations of Spalart-Allmaras turbulence model.
- Single species with passive scalars for visualisation.
- Single and multiple moving reference frame formulation.
- Multicomponent flows using the diffused interface model.
Spatial Discretisation
- 1st–order Upwind for fast solutions.
- MUSCL 2nd- and high-order scheme, with a wide choice of limiters.
- Central, WENO, CWENO, CWENOZ schemes up to 7th-order of accuracy.
- Reconstruction wrt conservative, primitive, characteristic variables.
- Generic, Legendre, Taylor polynomials for k-exact reconstruction.
- Modal Discontinuous-Galerkin discretisation.
- Symmetric Gaussian quadrature rules.
- Green Gauss and least-square gradient approximations.
- Low-Mach number correction for low speed flows.
- Multi-Dimensional Optimal Order Detection (MOOD) algorithm.
Riemann Solvers
- HLLC.
- HLL.
- Rusanov.
- Roe.
- Hybrid Roe-Rusanov, HLLC-HLL
Time Discretisation
- Explicit 1st-order Euler.
- 2nd- to 4th-order Runge Kutta.
- Implicit 1st-order with Jacobi and LU-SGS for steady state problems.
- Dual Time Stepping with 2nd-order Implicit (Jacobi and LU-SGS) for unsteady problems.
- Dual Time Stepping with 2nd-order Explicit time stepping for unsteady problems.
Architecture
- Fortran 2003 OOP.
- MPI only and MPI+OpenMP modes.
- Intel MKL or OpenBLAS libraries.
- ParMETIS integration for mesh partitioning.
I/O
- Input in Ansys Fluent *.msh files, STAR-CD *.cel, *.vrt, *.bnd files, and Ugrid formats.
- A single output from all processors in ASCII or binary format Tecplot (embedded libraries) and VTK Paraview formats.
- Fully parametrisable/customised solution initialisation and boundary conditions for various problems.
- A single checkpoint file that can be used by any number of CPUs.
- Averaging within the UCNS3D code for unsteady simulations where the Reynolds stresses, and average values are written in Tecplot/Paraview output files.
- Probe positions with unsteady data at selected locations.
Features to be included in upcoming releases
- Sliding mesh interface.
- Adaptive mesh refinements.
- Additional turbulence models.
Want to see more features ?
Let us know now, get in touch ucns3d@gmail.com
© 2023 UCNS3D team