Semi-empirical Modeling To Predict Radiation Detection Performance For Dynamic Nuclear Security Scenarios

To fully evaluate the performance of radiation detection systems in nuclear security use-case specific scenarios requires time and resource intensive test campaigns. Modeling and simulation tools help to reduce this burden; however, the available tools are limited when testing scenarios where relative motion between the radiation source and the detector system exists. If the radiation detection system under test contains a proprietary nuclide identification algorithm, then this further limits the usefulness of currently available software. On behalf of the National Nuclear Security Administration Office of Nuclear Smuggling Detection and Deterrence, Los Alamos National Laboratory has developed a tool known as Detector Response for In-motion Virtual Experiments (DRIVE) that relies on a small set of measured data to produce use-case specific, time-series, spectral gamma detector, and gross count neutron detector responses in formats that are compatible with vendor specific identification algorithms. The tool offers immense flexibility in the scenarios that can be rapidly simulated to offer order-of-magnitude estimates on the detection performance for the system. The adjustable parameters include relative speed and distance between source and detector, source strength scaling, background radiation scaling, number and location of detector modules, the ability to inject multiple source signatures in a single configuration, background suppression, background variation, sample rate, and number of trials to create (with statistical variation) for a given configuration. All parameters are modified through an easy-to-use user interface. The paper presents the methodology for developing this tool, its benefits and limitations, and assumptions, along with several benchmark comparisons to real-world testing.