Introduction
The Launch, Ascent, and Vehicle Aerodynamics (LAVA) Framework is a suite of Computational Fluid Dynamics (CFD) and multi-physics solvers with accompanying toolsets for pre- and post-processing including grid generation.
With origins in the late 2000's, LAVA has been continually developed in the Computational Aerosciences branch of NASA Ames Research Center to address a wide variety of NASA's fluids modeling and simulation needs. The solver is built on a modern and flexible software architecture with extensive continuous integration testing at scale, pushing the state-of-the-art in both CFD methods and high-performance computing.
LAVA AMS Seminar
Applications
Some notable areas of application have been:
- Launch pad pressure and thermal environment analysis, including multi-phase simulations of plume-water interactions, over-pressure wave dynamics, and conjugate heat transfer for launch pad components. These capabilities have been applied to Space Shuttle, Artemis, and commercial launch vehicles.
- Aerodynamic performance characterization and sonic boom predictions for the X-59.
- Atmospheric re-entry vehicle deceleration simulations, including supersonic parachute deployment with fluid-structure interaction, dynamic stability analysis with 6DOF motion, and supersonic retro-propulsion dynamics.
- Aero-acoustic predictions for propulsion systems, including open-rotor noise analysis and X-59 jet noise assessment.
- Aerodynamic buffeting mitigation strategies for Artemis at the solid rocket booster forward attachment point.
- Conceptual design evaluation and optimization of novel airframe concepts through programs such as SUSAN and HEATHeR.
- High-fidelity scale-resolving simulations of aircraft operating with deployed high-lift systems during takeoff and landing phases.
Please see the Applications Gallery for more details on some of these applications.
Solver Capabilities
The development of LAVA's capabilities is guided by the application needs. To enable large-scale simulations with relatively modest compute resources and provide the most efficient and appropriate grid paradigms for each application, LAVA emphasizes both algorithmic innovation and code optimization for high-performance computing environments.
A central philosophy of LAVA is to reduce user-effort and improve error-prone aspects of the CFD workflow through extensive automation and robust tooling. This is exemplified by LAVA's support of multiple meshing paradigms: block-structured curvilinear with overset, Cartesian immersed boundary with Adaptive Mesh Refinement (AMR), and arbitrary polyhedral unstructured meshes. These methods provide users with the flexibility to choose the appropriate mesh paradigm for their specific application, ranging from fully hand-tailored body-fitted curvilinear meshes to partial automation using hybrid curvilinear/immersed-boundary approaches, to fully automated Cartesian or Voronoi unstructured meshes. Additionally, LAVA provides comprehensive in-situ and post-processing tools to streamline the entire simulation workflow.
A high-level overview of LAVA solver capabilities is given below:
| Capability | Cartesian | Curvilinear | Unstructured |
|---|---|---|---|
| Mesh type | Cartesian | Block-structured | Arbitrary polyhedral unstructured |
| Equations | Navier-Stokes | Navier-Stokes | Navier-Stokes |
| Discretization | Finite difference | Finite difference | Finite volume |
| Compressibility | Compressible | Compressible | Compressible |
| Fluid Model | Perfect, Stiffened | Perfect | Perfect |
| Boundary representation | Immersed Boundary | Body-fitted/immersed boundary | Body-fitted |
| Mesh transitions | Octree 2:1 interfaces | Overset or multi-block | Conformal |
| Spatial order of accuracy | 2nd, 4th | 2nd, 4th | 2nd |
| Temporal order of accuracy (explicit) | 2nd, 3rd, 4th | 3rd | 2nd, 3rd, 4th |
| Temporal order of accuracy (implicit) | ✗ | 2nd | 2nd |
| Solution Approaches | WMLES | WMLES/HRLES/RANS | WMLES/RANS |
| Mesh Adaptation | Adaptive mesh refinement | Adjoint based redistribution | ✗ |
| Moving bodies | Prescribed/6DOF motion | Prescribed | ✗ |
| Multi-species | ✓ | ✗ | ✓ |
| Multi-phase | ✓ | ✗ | ✗ |
| Reacting flows | ✗ | ✗ | ✗ |
| Radiative heat transfer | ✗ | ✗ | ✗ |
| Fluid-structure interaction | Shell element | ✗ | ✗ |
| Conjugate heat transfer | ✗ | ✗ | Coupled 1-D solid conduction |
| Shape optimization | ✗ | ✓ | ✗ |
| Hardware | CPU, static geometry on GPU | CPU, WMLES on GPU | CPU, WMLES on GPU |
RANS: Reynolds-averaged Navier-Stokes
HRLES: Hybrid Reynolds-averaged Navier-Stokes large-eddy simulation
WMLES: Wall-modeled large-eddy simulation