One of the inherent risks involved in testing or operating a rocket engine is the possible release of propellant, leading to an explosion. Potential scenarios may range from a small cloud of fuel vapor expelled from an engine prior to ignition during normal operations, to an accident or launch abort event where a large portion of the available propellant is released. Understanding the explosion mechanisms and blast environment is critical to designing systems that maximize personnel and facility safety.
Historically, the strength of a blast wave has been calculated using semi-empirical models, which provide pressure as a function of radial distance from the blast center. In the presence of a complex structure, such as a launch pad or test facility, secondary waves can be generated through reflection and diffraction, which can produce even higher pressures than the initial blast wave. In these cases, semi-empirical models are no longer sufficient, and computational fluid dynamics (CFD) simulations are used to provide higher-fidelity predictions of the blast waves and their interactions.
As part of a NASA Engineering & Safety Center assessment of CFD tools for blast prediction, we used the Loci/BLAST software to simulate the detonation of a high explosive as a surrogate for a propellant vapor cloud at a NASA rocket engine test facility. To resolve the complex test facility geometry within a 60 foot radius, these time-accurate simulations were performed on a computational mesh containing 270 million cells. The simulations results were validated with data gathered from detonation tests performed on site.
Following the success of these simulations, we have begun to model explosions of the actual propellant vapor cloud using finite-rate chemistry for the combustion of the fuel. This additional fidelity is important in applications in which the region of interest is close to the explosion, where the blast wave from a vapor cloud will generally travel slower and produce lower pressures than a high explosive blast.
Results and Impact
Visualizations of our blast simulations have provided significant insight into the propagation of blast waves and their subsequent interactions with the test facility structure. Results show peak pressures from secondary waves in excess of the initial blast wave. These results have been validated with test data in terms of both blast arrival time and magnitude.
The methodology developed in this work is being further refined to characterize blast environments for NASA’s Space Launch System. These improved blast prediction capabilities will enable critical vehicle and crew risks to be better defined and mitigated.
Role of High-End Computing Resources
These blast environment simulations are very computationally demanding, due to the inclusion of finite-rate chemistry calculations and the large computational meshes required to resolve the complex geometries involved. NASA supercomputing resources have been vital to completing this work. Each simulation was performed on the Pleiades supercomputer using approximately 2,000 processors, and generated over 2 terabytes of data.Brandon Williams, NASA Marshall Space Flight Center