Sean M. Robinson

Statistical methods applied to gamma-ray spectroscopy algorithms in nuclear security missions

Gamma-ray spectroscopy is a critical research and development priority to a range of nuclear security missions, specifically the interdiction of special nuclear material involving the detection and identification of gamma-ray sources. We categorize existing methods by the statistical methods on which they rely and identify methods that have yet to be considered. Current methods estimate the effect of counting uncertainty but in many cases do not address larger sources of decision uncertainty, which may be significantly more complex. Thus, significantly improving algorithm performance may require greater coupling between the problem physics that drives data acquisition and statistical methods that analyze such data. Untapped statistical methods, such as Bayes Modeling Averaging and hierarchical and empirical Bayes methods, could reduce decision uncertainty by rigorously and comprehensively incorporating all sources of uncertainty. Application of such methods should further meet the needs of nuclear security missions by improving upon the existing numerical infrastructure for which these analyses have not been conducted.

Time Series Evaluation of Radiation Portal Monitor Data for Point Source Detection

The time series of data from a radiation portal monitor (RPM) system are evaluated for the presence of point sources by isolating the contribution of anomalous radiation. Energy-windowed background spectra taken from the RPM are compared with the observed spectra at each time step during a vehicle drive-through. The total signal is turned into a ¿spectral distance¿ index using this method. This provides a time series with reduced systematic fluctuations due to background attenuation by the vehicle, and allows for point source detection by time-series analyses. The anomalous time series is reanalyzed by using a wavelet filter function of similar size to the expected source profile. A number of real drive-through data sets taken at a U.S. port of entry are analyzed in this way. A set of isotopes are injected into the data set, and the resultant ¿benign¿ and ¿injected¿ data sets are analyzed with gross-counting, spectral-ratio, and time-based algorithms. Spectral and time methods together offer a significant increase to detection performance.

Final Technical Report for the Neutron Detection without Helium-3 Project

This report details the results of the research and development work accomplished for the ‘Neutron Detection without Helium-3’ project conducted during the 2011-2013 fiscal years. The primary focus of the project was to investigate commercially available technologies that might be used in safeguards applications in the relatively near term. Other technologies that are being developed may be more applicable in the future, but were outside the scope of this study.

Optimal Background Attenuation for Fielded Radiation Detection Systems

Radiation detectors are often placed in positions difficult to shield from the effects of terrestrial background. This is particularly true in the case of Radiation Portal Monitor (RPM) systems, as their wide viewing angle and outdoor installations make them susceptible to terrestrial background from the surrounding area. A low background is desired in most cases, especially when the background noise is of comparable strength to the signal of interest. The problem of shielding a generalized RPM from terrestrial background is considered. Various detector and shielding scenarios are modeled with the Monte-Carlo N Particle (MCNP) computer code. Amounts of nominal-density shielding needed to attenuate the terrestrial background to varying degrees are given, along with optimal shielding geometry to be used in areas where natural shielding is limited, and where radiation detection must occur in the presence of natural background. Common shielding solutions such as steel plating are evaluated based on the signal to noise ratio and the benefits are weighed against the incremental cost.

Non-invasive material discrimination using spectral x-ray radiography

Current radiographic methods are limited in their ability to determine the presence of nuclear materials in containers or composite objects. A central problem is the inability to distinguish the attenuation pattern of high-density metals from those with a greater thickness of a less dense material. Here, we show that spectrally sensitive detectors can be used to discriminate plutonium from multiple layers of other materials using a single-view radiograph. An inverse algorithm with adaptive regularization is used. The algorithm can determine the presence of plutonium in simulated radiographs with a mass resolution per unit area of at least 0.07 g cm−2.

A Simulation Framework for Evaluating Detector Performance in Cargo Screening Applications

Deployed radiation portal monitor systems (RPMs) screen gamma-ray signatures of cargo at international border crossings with the goal of detecting illicit radiological materials. Estimating the detection sensitivity of these systems requires an in-depth understanding, and quantification, of RPM response to both benign and illicit sources. Benign sources of radioactivity include background, alterations of the background due to the presence of vehicles and cargo, as well as sources that frequently cause nuisance alarms. These nuisance sources, for example those consisting of naturally occurring radioactive materials (NORM) and medical isotopes, frequently limit system performance. Advanced detector technology promises to increase the capability of deployed systems to discriminate illicit from nuisance sources. Presented here is a framework developed to assess the performance of these passive detection technologies. Due to the difficulty in obtaining empirical data for emerging technologies, the foundation of this comparison framework lies on a simulated benign source population to create a comprehensive set of data representing cargo vehicles driving through the RPM. Quantification of performance stems from injecting simulated signatures from illicit sources and comparing probabilities of detection via case and population studies.

Improvements in the method of radiation anomaly detection by spectral comparison ratios.

We present a new procedure for configuring the Nuisance-rejection Spectral Comparison Ratio Anomaly Detection (N-SCRAD) method. The procedure minimizes detectable count rates of source spectra at a specified false positive rate using simulated annealing. We also present a new method for correcting the estimates of background variability used in N-SCRAD to current conditions of the total count rate. The correction lowers detection thresholds for a specified false positive rate, enabling greater sensitivity to targets.

Characteristics of multiprocessing MCNP5 on small personal computer clusters

The feasibility and efficiency of performing MCNP5 calculations with a small, heterogeneous computing cluster built from Microsoft® WindowsTM personal computers (PC) are explored. The performance increases that may be expected with such clusters are estimated for cases that typify general radiation-shielding calculations. Our results show that the speed increase from additional slave PCs is nearly linear up to 10 processors. Guidance is given as to the specific advantages of changing various parameters present in the system. Implementing load balancing, and reducing the overhead from the MCNP rendezvous mechanism add to heterogeneous cluster efficiency. Hyper-threading technology and matching the total number of slave processes to the total number of logical processors also yield modest speed increases in the range below 7 processors. Because of the ease of acquisition of heterogeneous desktop computers, and the peak in efficiency at the level of a few physical processors, a strong case is made for the use of small clusters as a tool for producing MCNP5 calculations rapidly, and detailed instructions for constructing such clusters are provided.

Combining Radiography and Passive Measurements for Radiological Threat Localization in Cargo

Detecting shielded special nuclear material (SNM) in a cargo container is a difficult problem, since shielding reduces the amount of radiation escaping the container. Radiography provides information that is complementary to that provided by passive gamma-ray detection systems: while not directly sensitive to radiological materials, radiography can reveal highly shielded regions that may mask a passive radiological signal. Combining these measurements has the potential to improve SNM detection, either through improved sensitivity or by providing a solution to the inverse problem to estimate source properties (strength and location). We present a data-fusion method that uses a radiograph to provide an estimate of the radiation-transport environment for gamma rays from potential sources. This approach makes quantitative use of radiographic images without relying on image interpretation, and results in a probabilistic description of likely source locations and strengths. We present results for this method for a modeled test case of a cargo container passing through a plastic-scintillator-based radiation portal monitor and a transmission-radiography system. We find that a radiograph-based inversion scheme allows for localization of a low-noise source placed randomly within the test container to within 40 cm, compared to 70 cm for triangulation alone, while strength estimation accuracy is improved by a factor of six. Improvements are seen in regions of both high and low shielding, but are most pronounced in highly shielded regions. The approach proposed here combines transmission and emission data in a manner that has not been explored in the cargo-screening literature, advancing the ability to accurately describe a hidden source based on currently-available instrumentation.

The Effect of the Three-Dimensional Geometry of Cargo on the Detection of Radioactive Sources in Cargo Containers

Pacific Northwest National Laboratory has developed computer models to simulate the screening of vehicles and cargo with radiation portal monitors for the presence of illegitimate radioactive material. In addition, selected measurements have been conducted to validate the models. An important consideration in the modeling of realistic scenarios is the influence of the three-dimensional geometry of the cargo on the measured signature. This is particularly important for scenarios where the source and detector move with respect to each other. Two cases of the influence of the three-dimensional geometry of the cargo on the measured radiation signature are analyzed. In the first, measurements show that spectral data collected from moving sources so as to maximize the gross-counting signal-to-noise ratio has minimal spectral distortion, so that the spectral data can be summed over this time interval. In the second, modeling demonstrates that the ability to detect radioactive sources at all locations in a container full of cargo scales approximately linearly with the vertical height of the detector, suggesting that detectors should be approximately the same height as the container they scan.