BLIZZARD

Background

Simulating an atmospheric re-entry is a complex challenge due to the multiphysics involved, which currently makes high-fidelity numerical simulation (CFD) codes unusable throughout the trajectory due to their excessive computation time. In addition, these methods require the generation of a high-quality mesh to accurately capture the gradients in the boundary layer. However, this task requires human intervention, which can take several days for a single geometry.

In the context of an atmospheric re-entry of a moving or deforming object or other structure, these calculations must be performed throughout the trajectory, requiring frequent re-meshing.

Challenge & Solution

To assess the risk posed by debris on the ground, various rapid-response tools have been developed by space agencies, research institutes and the private sector. These reduced models have been designed to calculate the distributions of pressure coefficients and wall heat fluxes at each instant of atmospheric re-entry, in order to determine the trajectory, thermal degradation and fragmentation of debris as realistically as possible.

However, these models do not take into account the physical complexity required to accurately predict structural degradation for certain geometries (such as concave shapes) or when more complex physical phenomena, such as shock-shock and shock-structure interactions, occur.

The BLIZZARD module was created to enable fully automatic meshing and supersonic calculation from a geometry in less than 5 minutes for suitable refinement.

Calculation Examples

Graphical Interface

A graphical interface has been developed in C++ with Qt5 to enable the calculation to be prepared, launched and post-processed, and the progress of the results to be monitored in real time.

Web Service

A web service developed in JavaScript is also available, free of charge to beta testers. Visit solve.rtech.fr to launch and automatically post-process CFD calculations via the web application.

Operating Principle

The software is based on the solution of the Euler equations using a Riemann solver coupled with the automatic generation of an octree Cartesian mesh (Aftosmis, 1997).

Mesh Generation

The method consists of generating a Cartesian mesh of the domain and truncating the cells on the surface or in the structure under consideration. The surface will then be represented by a set of cubes, creating a "stair-step" geometry boundary.

Figure 1: Representation of the creation of a mesh by an octree algorithm for a circle

Octree Architecture

The octree mesh uses binary mapping to store connectivity information using a binary tree system. The leaves of the tree represent the cells. Each child is linked by a pointer to a parent cell. This architecture enables efficient mesh refinement and connectivity management.

Figure 2: Representation of the creation of a 2D octree (quadtree)