The Reacting Flow Environments branch at NASA ARC is interested in characterizing the aerothermal environment of three main classes of problem: planetary entry vehicles, reusable launch vehicles (RLVs), and arc-jet (or other ground test) flow simulations. Each of these problem classes has unique physical characteristics, the understanding of which is at the cutting edge of the field. Proper modeling of the relevant physics is required to accurately simulate the aerothermal environments of these problem classes. These include, but are not limited to, chemical non-equilibrium, thermal non-equilibrium, shock layer radiation, surface catalycity, and thermal protection system material interaction with the aerothermal environment.
The primary tool for solving this class of problems has been commercial and government codes such as GASP and LAURA. Existing CFD (computational fluid dynamics) codes such as GASP are not ideally suitable. The Data Parallel Line Relaxation (DPLR) software package is a suite of CFD tools for the computation of supersonic and hypersonic flows in chemical and thermal non-equilibrium. Included in the package are 2D/axisymmetric and 3D structured grid finite volume Navier-Stokes codes, a pre-processor, and a post-processor. The CFD solver is written in Fortran 90 and supports distributed memory parallelism through MPI. The code supports implicit boundary conditions, generalized multi-block topologies, grid alignment to flow features, and generalized chemical kinetics and thermodynamic property databases. Physical modeling is sufficient to solve perfect gas, dissociating, and weakly ionized flows in various states of non-equilibrium. DPLR is currently the primary aerothermal analysis tool at NASA Ames and Johnson Space Center, and is used at two other NASA centers, multiple companies, and several universities in support of all four NASA mission directorates; AFRL (Air Force Research Laboratory), AFOSR (Air Force Office of Scientific Research), and DARPA (Defense Advanced Research Projects Agency) research.
The DPLR package incorporates several pieces of existing code into its architecture:
- Tecplot I/O libraries (optional). Linking to these libraries enables post-processing data output directly in a binary format native to the Tecplot post-processing software. This library is provided with each licensed Tecplot distribution and can be linked in at compile time if available.
- FXDR libraries (optional). The use of XDR binary format for grid and restart file I/O permits architecture independence of these files. In other words, a problem initially run on Origin can be restarted or post-processed on an Alpha machine, completely transparently to the user.
- MPI (required). DPLR is inherently an MPI code, and cannot run without some form of MPI installed on the user’s machine. There are many commercial versions of MPI available for nearly all computer architectures. In addition, a freeware version of MPI, called MPICH, is available in a source distribution from the Argonne National Laboratory at https://www.unix.mcs.anl.gov/mpi/mpich/indexold.html.
- BLAS libraries (optional). Linking to architecture-specific BLAS libraries can increase performance by about 25% on many computer architectures.
Novel features of DPLR include code generality, sophisticated physical modeling, fully pointwise and implicit boundary conditions, extensive post-processing and data extraction capabilities, high parallel efficiency, and scalability. In addition, the code has the capability of automatically detecting block interfaces and tailoring the volume grid to the hypersonic shock layer. Extensive support for advanced TPS boundary conditions, such as generalized catalysis and radiative equilibrium walls, is also supported. New boundary conditions are provided to facilitate the solution of smaller unit problems within a larger construct. The code also provides APIs for loose coupling to other shock layer radiation and material response solvers. Finally, the code supports virtual domain decomposition of multiblocked structured grids, which greatly simplifies problem setup.