- R Paul Drake
- The specific focus of the project was radiative shocks, which develop when shock waves become so fast and hot that the radiation from the shocked matter dominates the energy transport. This in turn leads to changes in the shock structure. Radiative shocks are challenging to simulate, as they include phenomena on a range of spatial and temporal scales and involve two types of nonlinear physics Ð- hydrodynamics and radiation transport. Even so, the range of physics involved is narrow enough that one can hope to model all of it with sufficient fidelity to reproduce the data. CRASH was focused on developing predictions for a sequence of experiments performed in Project Year 5, in which those experiments represented an extrapolation from all previously available data. The previous data involved driving radiative shocks within cylindrical structures, and mainly straight tubes. The Year 5 experiments drove a radiative shock down an elliptical tube. Our long-stated goal for these predictions was that the distribution of predicted values would overlap significantly with the observed distribution. We achieved this goal. Achieving our goal required the conversion of an established space-weather code to model radiative shocks at high energy density. To obtain reasonable fidelity with respect to the experimental data required implementing a laser absorption package, in addition to a hydrodynamic solver, electron physics and heat conduction, and multigroup diffusive radiation transport. The dedicated experiments provided evidence of experimental variability, validation of the calculation of initial shock wave behavior, and validation data at many observation times using cylindrical shock tubes. Following this were preparatory experiments for and finally the execution of the Year 5 experiments. The predictive science research included a wide range of sensitivity studies to determine which variables were important and a sequence of predictive studies focused on specific issues and sets of data. This led ultimately to predictions of shock location for the Year 5 experiments. A conclusion from this project is that the serious quantification of uncertainty in simulations is a dauntingly difficult and expensive prospect. Pre-existing codes are unlikely to have been built with attention to what will be needed to quantify their uncertainty. Pre-existing experimental results are even more unlikely to include a sufficiently detailed analysis of the experimental uncertainties. And this will also be true of most experiments that might be used to validate components of the simulation. The analysis of uncertainty in any one of the physical processes (and related physical constants) is a major effort. And addressing model form uncertainty is an even bigger challenge, that may in principle require development of complete, alternative simulation models. We made a start at all of this, and completed almost none of it. But by the end of a project, we finally had all the pieces in place and working that would have enabled a range of important studies and advances in relatively near-term years. But the sponsor terminated the program after only five years. For most of the participants this was a relatively minor development, although for a few of them it proved to be enormously disruptive. We believe that the cost to the nation, in work that was ready be done but now will not be, was much much larger. The sketch of the target was produced using a drawing program based on the experimental dimensions. The annotated photograph of the target was obtained using a visible-light camera. The colorized radiographs were obtained via backilit-pinhole radiography of a radiative shock propagating down an elliptical tube, at 26 ns after the lasers driving the shock tube fired. The graph showing lines and circles was produced by running many computer models, analyzing their statistical distribution, and measuring actual shock positions in the experiment.
- Radiative shock