Engineering topics

Gerts, David

Publications and source records attributed to Gerts, David.

3 records

DOE OSTI2010

A Transferrable Belief Model Representation for Physical Security of Nuclear Materials

This work analyzed various probabilistic methods such as classic statistics, Bayesian inference, possibilistic theory, and Dempster-Shafer theory of belief functions for the potential insight offered into the physical security of nuclear materials as well as more broad application to nuclear non-proliferation automated decision making theory. A review of the fundamental heuristic and basic limitations of each of these methods suggested that the Dempster-Shafer theory of belief functions may offer significant capability. Further examination of the various interpretations of Dempster-Shafer theory, such as random set, generalized Bayesian, and upper/lower probability demonstrate some limitations. Compared to the other heuristics, the transferrable belief model (TBM), one of the leading interpretations of Dempster-Shafer theory, can improve the automated detection of the violation of physical security using sensors and human judgment. The improvement is shown to give a significant heuristic advantage over other probabilistic options by demonstrating significant successes for several classic gedanken experiments.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2010

Nuclear Material Accountability Applications of a Continuous Energy and Direction Gamma Ray Detector

The Idaho National Laboratory has recently developed a detector system based on the principle of a Wilson cloud chamber that gives the original energy and direction to a gamma ray source. This detector has the properties that the energy resolution is continuous and the direction to the source can be resolved to desired fidelity. Furthermore, the detector has low power requirements, is durable, operates in widely varying environments, and is relatively cheap to produce. This detector is expected, however, to require significant time to perform measurements. To mitigate the significant time for measurements, the detector is expected to scale to very large sizes with a linear increase in cost. For example, the proof of principle detector is approximately 30,000 cm3. This work describes the technical results that lead to these assertions. Finally, the applications of this detector are described in the context of nuclear material accountability.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2010

Using GPU Programming for Inverse Spectroscopy

The Idaho National Laboratory (INL) has developed a detector that relies heavily on computationally expensive inverse spectroscopy algorithms to determine probabilistic three dimensional mappings of the source and its intensity. This inverse spectroscopy algorithm applies to material accountability due to the potential to determine where nuclear sources are present as a function of time and space. And yet because the novel algorithm can become prohibitively expensive on a standard desktop PC, the INL has incorporated new hardware from the commercial graphics community. General programming for graphics processing units (GPUs) is not a new concept. However, the application of GPUs to evidence theory-based inverse spectroscopy is both novel and particularly apropos. Improvements while using a (slightly upgraded) standard PC are approximately three orders of magnitude, making a ten hour computation in less than four seconds. This significantly changes the concept of prohibitively expensive calculations and makes application to materials accountability possible in near real time. Indeed, the sensor collection time is now expected to dominate the time required to determine the source and its intensity, instead of the inverse spectroscopy method.

97 MATHEMATICS AND COMPUTING↗