Please note! All times in the online program are given in New York - America (GMT -04:00) times!

New York - America ()
Jan 29, 2022, 8:57:46 AM
Your time ()
Tokyo - Asia ()
Jan 29, 2022, 10:57:46 PM
Our exhibitors and sponsors – click on name to visit booth:

To search for a specific ID please enter the hash sign followed by the ID number (e.g. #123).

Computer Modeling II

Session chair: Shimazoe , Kenji (University of Tokyo, Department of Nuclear Engineering and Management, Bunkyo, Japan); Danagoulian , A (MIT, USA)
Shortcut: N-21
Date: Thursday, 21 October, 2021, 7:00 AM - 8:45 AM
Room: NSS - 1
Session type: NSS Session


Click on an contribution to preview the abstract content.

7:00 AM N-21-01

Feasibility study of linac-based Delayed Gamma Neutron Activation Analysis for copper characterization in scrap metal (#146)

R. De Stefano1, A. Sari1, H. Makil1, F. Carrel1, P. Russo2

1 Université Paris-Saclay, CEA LIST, F- 91120 Palaiseau, France
2 Arcelor Mittal, Global Research and Development, F-57283 Maizières-les-Metz, France


In the framework of its activities related to copper quantification in scrap metal, CEA LIST and ArcelorMital R&D investigate the use of a novel linac-based Neutron Activation Analysis (NAA) technique using a 9 MeV LINAC (Linear electron accelerator) coupled to a (γ, n) conversion target. As this method deploys a photon-interrogating source, it is possible to couple NAA with a primary X-ray radiography for mapping the heterogeneity of the investigated samples. This technique also allows high-level neutron emissions of the order of 1010 n.s-1. A former experimental study that was focused on the feasibility of detecting the 1039 keV delayed gamma-ray of 65Cu samples in several positions of a scrap metal mockup drum put to light an important signal dependency with regard to copper location and to its core geometrical characteristics. To interpret this dependency due mainly to self-protection, self-attenuation, and attenuation of the signal in the surrounding matrix, the study presented in this paper focuses on the activation of copper in a simple scenario with a scrap metal planar matrix. Several 65Cu distributions of 126.6 g are computed using the recent Activation Control card (ACT) of Monte Carlo code MCNP6.1, which allows the transport of activation-induced delayed gamma-rays in single-step calculations. One copper distribution is described as homogeneous, a second as a disc-shaped mass computed for five positions. The homogeneous distribution shows a significant increase of signal of more than a factor 3 with respect to the configuration with the solid mass in the middle of the metal matrix. We also show that the delayed gamma-ray signal decreases following a quadratic dependency with respect to the distance to the planar metal matrix center. We concluded that in this case scenario an ideal detector should scan a 12 × 12 cm2 surface to quantify 65Cu mass in scrap metal with a 10 % uncertainty.

Keywords: Delayed gamma-rays, linear electron accelerator, Neutron Activation Analysis, copper characterization
7:15 AM N-21-02

Grasshopper, a Geant4 front end: validation and benchmarking (#1256)

A. Danagoulian1, E. A. Klein1, J. N. Miske1

1 Massachusetts Institute of Technology, Nuclear Science and Engineering, CAMBRIDGE, Massachusetts, United States of America


Over the last two decades the Geant4 object-oriented Monte Carlo (MC) simulation toolkit has expanded in its application from high energy physics to multiple additional fields.  Geant4 has a number of advantages over other tools, however it requires considerable knowledge of the C++ programming language, as well as a thorough understanding of Geant4's own libraries and internal procedures.  These prerequisites make Geant4 a difficult choice for some research and educational projects.  To address this issue we have developed a new tool, grasshopper, which is an open source front end for Geant4.  It accepts XML input through the built-in Geant4 GDML parser, allows the user to specify source distributions, and enables rapid learning and fast setup of simulations.  In this paper we describe grasshopper, and provide basic benchmarking for a number of nuclear processes as well as validation against prior experimental data.  Grasshopper combines Geant4's power with the convenience of a simple interface that requires O(hr) to learn and O(10 min) to setup a simulation.  This would allow for a significant expansion of Geant4's applicability in various sub-fields of nuclear science, both in research and educational settings.

AcknowledgmentThis work was supported in part by Department of Energy Award No. DE-NA0003920, as part of the NNSA Monitoring, Verification, and Technology (MTV) consortium.  J.N. Miske gratefully acknowledges support from MIT's Undergraduate Research Opportunities (UROP) program.  E.A. Klein is supported by a Nuclear Nonproliferation International Safeguards (NNIS) Graduate Fellowship sponsored by the Department of Energy's National Nuclear Security Administration.
Keywords: Monte Carlo, Geant4, simulation
7:30 AM N-21-03

Development and Validation of a Metaheuristic Optimization Method for Neutron Spectra Tailoring at the National Ignition Facility. (#1458)

S. Bogetic1, L. Dauffy2, C. Yeamans2, J. Vujic3, D. Shaughnessy2

1 University of Tennessee, Department of Nuclear Engineering, Knoxville, Tennessee, United States of America
2 Lawrence Livermore National Laboratory, Nuclear and Chemical Sciences Division, Livermore, California, United States of America
3 University of California, Department of Nuclear Engineering, Berkeley, California, United States of America


The National Ignition Facility (NIF) at Lawrence Livermore National Laboratory (LLNL) is a unique source of neutrons that uses laser inertial confinement to drive the deuterium-tritium fusion reaction. The reaction produces a very high flux output and a monoenergetic 14.06 MeV source peak. With such a strong neutron source, various possibilities are opening up for a list of applications by tailoring the neutron spectra. Neutrons with various energy ranges could be used for detector calibrations, for radiation damage to different materials, for cross section measurements for neutron activation studies at NIF. At this purpose, two new software packages were developed to optimize designs of energy tuning assemblies (ETAs) to produce desired neutron spectra: Gnowee uses metaheuristic optimization algorithm to sample the design space efficiently; COEUS couples Gnowee with a transport solver to automatically generate ETAs given a large set of constrains, changeable variables and objective spectrum. In order to set the software to be a proof-of-concept platform, for future designs of energy tuning assemblies, a series of validation experimental processes have been performed at NIF. The final goal is to look into the large nuclear data uncertainties of the modified neutron spectra which, in a full ETA design, are hardly identifiable. Thus, the idea is to "isolate" and emphasize specific energy channels for the validation of known data weaknesses at 14 MeV and below, and to set a diagnostic SNOUTs or HTOAD with easy to field single or partial stack-up materials. In this way it is possible to investigate the impact of the data of each material of interest for NIF stewardship, on the ability to model future ETA experiments.

AcknowledgmentThis work was performed under the auspices of the U.S Department of Energy by Lawrence Livermore National Laboratory under contract DE-AC52-07NA27344. Funding was supported by the LLNL Graduate Scholar Program.
Keywords: NIF, optimization, activation, MCNP, validation
7:45 AM N-21-04

Agile Path Planning for Radiation Source Searching with Aerial Drones (#1245)

Q. Zhou1, C. Redding2, H. Qi1, J. Hayward2

1 The University of Tennessee, Department of Electrical Engineering and Computer Science, Knoxville, Tennessee, United States of America
2 The University of Tennessee, Department of Nuclear Engineering, Knoxville, Tennessee, United States of America


Small unmanned aerial vehicles may allow for rapid radioactive source clearing in an indoor environment. Our work on agile path planning is able to effectively search for and detect radioactive sources in such an environment. Our proposed path planner may potentially find a quicker, more optimal, and collision free route when compared to the path planned without radiation count data. Our entire system and algorithms are implemented and demonstrated within the ROS Gazebo simulation environment with Cs-137 radiation source.

AcknowledgmentThis material is based on work supported in part by the Defense Threat Reduction Agency under grant number HDTRA 1-18-1-005 and in part by the Department of Energy National Nuclear Security Administration through the Nuclear Science and Security Consortium under Award Number(s) DE-NA0003180 and/or DE-NA0000979.
Keywords: Unmanned Aerial Vehicle, Radiation SourceSearch, Robot Operating System, Indoor Source Clearing
8:00 AM N-21-05

A Modeling Framework for Distributed Radioactivity Sensor Networks (#220)

D. Raji1, 2, R. Cooper2, J. Hayward1, T. Joshi2, M. Salathe2

1 University of Tennessee, Nuclear Engineering, Knoxville, Tennessee, United States of America
2 Lawrence Berkeley National Laboratory, Nuclear Science Division, Berkeley, California, United States of America


We present on an original framework for modeling a theoretical wireless sensor network for distributed radiation mapping over a wide area. The main thrusts of this effort are developing realistic source terms using atmospheric simulations, modeling the network nodes’ topology and measurement of radiation intensity, and developing network-wide algorithms including spatial reconstruction and pathfinding. Source term data generated is used by network algorithms as ground truth to evaluate network fitness for different parameterizations. Relevant parameters include spatial size of deposition, terrain variation, quantity of nodes, and node placement scheme among others. A selection of quantitative results of the algorithm analysis are exhibited along with qualitative sample outputs.

AcknowledgmentThis material is based upon work supported by the Defense Threat Reduction Agency under DTRA 13081-31571 and by the Department of Energy National Nuclear Security Administration through the Nuclear Science and Security Consortium under Award Number DE-NA0003180.
Keywords: Sensor networks, distributed radioactivity, modeling and simulation
8:15 AM N-21-06

Trained Deep Convolutional Neural Network for Attenuated Gamma-Ray Detection Using CdZnTeSe Spectrometer (#1171)

S. K. Chaudhuri1, J. W. Kleppinger1, K. Roy2, R. Panta2, F. Agostinelli2, A. Sheth2, U. N. Roy3, R. B. James3, K. C. Mandal1

1 University of South Carolina, Department of Electrical Engineering, Columbia, South Carolina, United States of America
2 University of South Carolina, Artificial Intelligence Institute (AIISC) of the Department of Computer Science and Engineering, Columbia, South Carolina, United States of America
3 Savannah River National Laboratory, Materials and Devices Division, Aiken, South Carolina, United States of America


CdZnTeSe (CZTS) is a novel wide-bandgap quaternary semiconductor for gamma-ray detection. Being a room-temperature detector material CZTS finds immense scope of application in the field of in-situ monitoring of spent nuclear fuel (SNF) storage casks. The idea of employing a radiation monitor inside a cask or storage canister is challenging because of extremely high radiation dosage and high temperatures. Instead, a portable CZTS spectrometer can be used from outside the storage casks to monitor any abrupt changes in the energy and intensity levels of the gamma rays emitted by the SNF.  However, monitoring SNF by detecting such gamma rays using conventional spectroscopic methods is unreliable due to factors like significant variation of radiological background from the immediate surroundings, attenuation in shielding materials, and unavailability of statistically reliable data. We report the design and development of a deep convolutional neural network (CNN) to recognize spectral features in a sequence of data obtained from a NI PCI 5122 fast digitizer connected to a pre-amplifier and CZTS detector assembly. The CNN is designed for multiclass and multilabel classification leading to accurate and precise determination of gamma-ray energies regardless of the fluctuations due to the attenuation in the shielding materials and presence of radiological backgrounds. We have used simulated detector-preamplifier data with various incident energies to train a CNN on the equivalent of 90 seconds worth of simulated data and validated it on 10 seconds worth of simulated data with high accuracy.

AcknowledgmentThe authors acknowledge partial financial support from the DOE Office of Nuclear Energy’s Nuclear Energy University Programs (NEUP), Grant No. DE-NE0008662.
Keywords: nuclear material safeguard, gamma-ray detection, CdZnTeSe (CZTS), machine learning, deep convoluted neural network (CNN)
8:30 AM N-21-07

Using Decision Trees to calculate the effective atomic number from the Rayleigh to Compton Scattering Ratio (#1319)

G. Gert1, C. Mattoon1, B. Beck1

1 Lawrence Livermore National Laboratory, Physical and Life Sciences Directorate, Livermore, California, United States of America


The Rayleigh to Compton scattering ratio (RC ratio) has proved to be useful alternative in the calculation of effective atomic numbers (Zeff) in regimes that are beyond the limits of detectability of the attenuation coefficient. The methods that are used in these calculations have however not kept up with the state of the art in the available nuclear data, nuclear data processing methods, computational tools and data processing methods. This project highlights how the current state of the art may be applied to the calculation of Zeff for the RC ratios. The FUDGE nuclear data processing code was used to process the ENDF/B-VIII.0 photo-atomic data to extract the coherent and incoherent scattering factors at a set of given incident photon energy and scattering angle. Principal Component Analysis (PCA) was then used to interrogate the complex relationship between the RC ratios and Z values. A Decision Tree Regression model was then used to predict Zeff values from the corresponding RC ratios. The preliminary results highlight the shortcomings in the training data and the current focus is to expand the data to a more representative set that is not just limited to the current integer Zeff values. 

AcknowledgmentThis work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory. Document release# LLNL-ABS-822315
Keywords: Effective Atomic Number, FUDGE, GNDS, Raleigh to Compton Scattering Ratio

Our exhibitors and sponsors – click on name to visit booth: