 |
|
|
|
The DOE weapons laboratories are responsible for identifying and conducting studies necessary to ensure confidence in the safety and performance of weapons in the U.S. nuclear stockpile. Significant increases in computing capabilities will be required to meet that responsibility in the future, due to the limitations (financial, political, and environmental) that are expected on all aspects of weapon systems testing.
In an unclassified context, it can be said that computational simulations of the nuclear explosive package itself involve tightly coupled models of hydrodynamics, radiation transport, and fission/fusion energy release. In the past, underground nuclear test data were used to normalize the computational models (adjust model parameters, of ``knobs'') and reduce the modeling uncertainties introduced by these approximations. Now, however, various approximations are required in the computational modeling for safety and performance predictions, due to limitations in current computing capabilities.
To offset the anticipated unavailability of new test data, current computational simulations for evaluating the safety or performance of the nuclear explosive package must be enhanced to provide (1) greater spatial resolution, (2) full three-dimensional representation of the device behavior and physical environments, and (3) more completeness and fidelity in the physical processes represented. One thrust of the DOE ASCI program is to develop this ``no-knobs'' enhanced modeling capability for safety and performance assessments of the nuclear device. Thousandfold increases in computer speed, memory, and storage capacity will be needed to support the improved simulations.
In addition to the nuclear explosive package, a nuclear weapon system involves a complex assembly of electrical and mechanical components, each of which is critical to some aspect of the safety, security, or performance of the system. Evaluation of the weapons system response to various hypothetical accident conditions presents a considerable challenge to the safety analyst, because of the the severity, complexity, and number of potential accident environments.
Full-scale testing for some accident scenarios is prohibitively expensive and, in some cases, could involve conditions that are hard to reproduce without causing adverse environmental impact. Thus, a virtual test capability is needed to provide a cost-effective and environmentally benign option fir assessing and advancing weapon safety. Such a virtual testing capability would involve high-fidelity computational models to simulate the environments, evaluate component response, assess system safety, and guide enhanced component designs for the spectrum of environments that may occur over the lifetime of a weapon system. confidence in the accuracy of the calculations will be developed through carefully controlled and instrumented tests, which will provide data for validating the physics modules in the simulation codes. Development of the high-performance computing technology to support a virtual testing and safety analysis capability for weapon nonnuclear components and subsystems is another goal of this ASCI program element.
One of the risk-dominant accident scenarios for nuclear weapons delivered by aircraft is the crash of the airplane on takeoff or landing with the weapon in carriage. In this scenario, it is assumed that the jet fuel catches fire as a result of the crash and engulfs the weapon for some prolonged period of time. the ``crash and burn'' scenario involves multiple, combined physical insults to the weapon system. In particular, severe mechanical deformations of the various nuclear and nonnuclear components can occur as the aircraft and weapon strike the the runway. Sustained exposure to the intense heat of the jet fuel fire following the crash causes decomposition of many of the nonmetallic materials in the nonnuclear components. Accurate analysis of this type of accident, therefore, requires a coupled thermal-mechanical-chemical modeling capability to simulate the response of critical reactive components, such as detonators, polymeric insulators, and other similar devices.
The computational tools for treating the coupled physics of weapon accident environments are still to be developed, though models for many of the individual physical effects do already exist. A thousandfold increase in computer speed and memory capabilities will be required for these calculational analyses, in order to adequately represent the complexity of the problem. This need for increased capability is driven by the need for increased resolution to capture details of component geometry and complex phenomenology, by the additional variables required for the coupled physics of the problem, and by the increased scale and complexity of modeling that will be required to simulate realistic, full-scale thermal and structural effects of the accident environment. it should also be noted that the computing capability requirements for safety analyses of the nonnuclear subsystems is comparable to that identified above for modeling the nuclear explosive package.
Stewardship of the nuclear stockpile is the primary mission of the DOE laboratories. Assuring safety of the weapons in the arsenal is a key element of that stewardship.
Weapon accident scenarios can involve complex combinations of physical environments. Since full-scale testing can be prohibitively expensive and hazardous to the environment, increased reliance will be placed on computational simulations to assess weapon safety.
New computational methods will be required to model complex weapon accident scenarios (see Figure 1 ). Furthermore, significant increases in computer capability will be required to support weapon safety analyses. An analysis of a risk-dominant aircraft crash environment will be the first goal of the program. Below, we list other roadmap estimates.
| [MCS | Announcement | Case Studies | Report | Attendees] | ||
|
