NERS Project #1: Radiation Weather Station (RWS)
Faculty Mentor: Professor Kim Kearfott, kearfott@umich.edu
Additional Mentor: Jordan D. Noey MS, noeyjd@umich.edu
Prerequisites: Students work as part of a team; a student’s specific assignment will depend upon the background and interests of each student
Project Description: Knowing changes in the presence of radionuclides in the environment is important for the detection and response to nuclear power plant accidents and radiological terrorist events. The Radiation Weather Station is a facility consisting of sensors located at several stations on the University of Michigan’s North Campus, with auxiliary stations planned elsewhere. Sensors are available for temperature, pressure, humidity, rainfall, wind speed and direction, solar radiation, solar flares, soil moisture, radon, and gamma rays. Naturally occurring, medical and nuclear power plant radionuclides may be detected, with information collected about the energy of their gamma rays. Monitoring is designed to identify radionuclides from natural, planned, and accidental releases while tracking indoor and outdoor environmental conditions. All data are shared through a continuously-updated website. A variety of projects are available involving the development and improvement of the database, deployment of new sensors, analysis of large data sets, website development, the addition of mobile phone-based radiation detectors, and the preparation of educational materials for the public and K-12 audiences. A small, affordable RWS featuring a smaller collection of more affordable sensors controlled by a Raspberry Pi computer is also undergoing development. Tasks and specific work are tailored to the student, depending upon their interests and capabilities. This may involve one (or possibly multiple) of the following: software usage, statistical analysis, coding, hardware interfacing, analysis of temporal data sets, machine learning, design and 3-D printing of mechanical parts and cases, design of circuits and printed circuit boards, historical research, and technical writing for the public.
Note: This project is *not* about climate change or weather prediction. It is about detecting and characterizing the type and source of ionizing radiation as it changes as a function of time at different locations.
Research Mode: remote, hybrid (some in-person may be possible/needed for certain tasks but not required of all participants). Example in-person tasks could include installation, modification, calibration, or testing of new sensors.
NERS Project #2: Smart Radiation Detectors and Adaptive Navigation for Radiation Surveys
Faculty Mentor: Professor Kim Kearfott, kearfott@umich.edu
Additional Mentor: Jordan D. Noey MS, noeyjd@umich.edu
Prerequisites: Students work as part of a team; a student’s specific assignment will depend upon the background and interests of each student
Project Description: Several software and hardware development projects are related to the improvement and testing of a smart radiation detector. The current design is a Geiger-Muller (GM) ionizing radiation detector controlled by Raspberry Pi computer. Enhanced designs may ultimately be possible using spectroscopic detectors. Smart radiation detectors could be used by individuals performing radiation surveys in smaller spaces such as individual laboratories. They could also be transported larger distances using ground-based or aerial platforms, ranging from firetrucks to automatically piloted drones. Software involving auditory or visual communications could alert surveyors of spots that have been missed in performing detailed surveys, such as lab benches when searching for contamination. More complex software could be used to construct the distribution of the sources of radiation in an environment from limited measurements. That approximation could then be used to inform the best locations for additional measurements and optimize a measurement path to further improve knowledge of the sources. Such approaches would be invaluable for first responders to radiological events as well as those involved in the cleanup of legacy radioactive wastes or the decommissioning of nuclear facilities. Experienced students are sought to develop computer programs and mobile device Apps to fully enable their appropriate functioning, display of data, and intelligent navigation of such detectors. Students should provide clear information about their experience and proficiency in software, hardware, radiation detection, and mechanical design so that their role on a diverse team working on this project could be identified.
Note: The algorithms in this project may also be applied to the location of WiFi emitters and other sources of radiation (such as radio waves) that are non-ionizing. Some projects are available concerned with the usage of these, which are “substitutes” for ionizing radiation.
Research Mode: remote, hybrid (some in-person may be possible/needed for certain tasks but not required of all participants). Dry benchwork (such as soldering involved in the making of a smart detector) may be completed at students’ homes and tested using readily available radioactive materials. Example in-person tasks could include experiments testing the performance of the smart detectors or algorithms for using them. Some experiments may be performed outside.
NERS Project #3: Radon Gas—An Indoor Radioactive Hazard Useful for the Detection of Nuclear Weapons and Earthquakes
Faculty Mentor: Professor Kim Kearfott, kearfott@umich.edu
Prerequisites: Students work as part of a team; a student’s specific assignment will depend upon the background and interests of each student
Project Description: Radon gas is a ubiquitous naturally occurring radioactive material that occurs throughout the environment and in all buildings, at least in small amounts. It can be readily detected but presents health hazards when in high concentrations. Radon gas levels change as a function of local weather conditions, as well as the heating or cooling situation within a building. Radon has also been observed to change many days in advance of major earthquakes. This project involves the study of radon gas as a function of time both indoors and outside. State-of-the-art equipment is deployed both to measure radon gas as well as to track local weather and other conditions such as solar and background radiation from other sources. Students may participate in both experimental data collection as well as analysis of large data sets. Discrimination of airborne transuranics from naturally occurring radon gas is particularly important for worker protection during commercial nuclear power plant outages and dose control during emergencies. The topic is especially suitable for students ultimately interested in homeland security/treaty verification, nuclear power plant operations, and/or radiation protection. Students should have motivation to learn, basic programming skills, and solid mathematics and physics backgrounds. Students with appropriate backgrounds may be able to apply machine learning techniques to analyze existing data sets.
Research Mode: remote, hybrid (some in-person may be possible/needed for certain tasks but not required of all participants). In-person tasks for some participants could include counting charcoal canisters or reading out electrets on existing laboratory equipment or setting up experiments to collect data.
NERS Project #4: Thermoluminescent and Optically Stimulated Luminescent Dosimetry– Radiation Dose Measurements for Workers, Patients, and Environmental Monitoring
Faculty Mentor: Professor Kim Kearfott, kearfott@umich.edu
Additional Mentor: Jordan D. Noey MS, noeyjd@umich.edu
Prerequisites: Students work as part of a team; a student’s specific assignment will depend upon the background and interests of each student
Project description: Dosimeters are passive, integrating materials used to monitor the radiation exposure of workers in nuclear facilities. Although all workers receive dosimeters, there are different types and they have different performance characteristics. New dosimeter types and ways of calibrating and deploying them are being developed in the laboratory. Dosimetry systems are also used for medical applications including radiation therapy, diagnostic radiology, and nuclear medicine. The limitations of different types of dosimeters are being actively compared and characterized for medical applications. Advanced software is also being developed to automatically analyze thermoluminescent dosimeter glow curves for research projects as well as routine analysis. A dosimetry calibration source is also being carefully characterized using quality control experiments undergoing development. Students may be engaged in performing experiments, data analysis, and/or software design. This project is especially suitable for students ultimately interested in homeland security/treaty verification, medical physics, nuclear power plant operations, and/or radiation protection. Students should have the motivation to learn, basic programming skills, and solid mathematics and physics or computer programming backgrounds.
Research Mode: remote, hybrid (some in-person may be possible/needed for certain tasks but not required of all participants). Examples: in-person tasks could include performing quality control of a radiation source, calibration of detectors, processing of dosimeters, or performance of experiments involving phantoms.
NERS Project #5: Radiation Spectroscopy for the Identification of Radionuclides
Faculty Mentor: Professor Kim Kearfott, kearfott@umich.edu
Additional Mentor: Jordan D. Noey MS, noeyjd@umich.edu
Prerequisites: Students should have strong skills, experience, or interest in computer programming or radiation detection
Project description: The applications of this project are the protection of the public from environmental radiation, nuclear accident dose reconstruction, and nuclear weapons treaty verification. Energy spectroscopy involves the determination of the energy of particular types of radiation, which are characteristic of the source of radiation. Alpha, gamma, and neutron spectroscopic devices are calibrated and deployed to solve real-world problems involving radiation sources. Students may become involved in nuanced calibrations, data interpretation, and specific measurement campaigns involving a variety of both state-of-the-art and newly developed instruments used for radiation spectroscopy. Applications of an imaging spectrometer to the medical environment as well as for naturally occurring radioactivity may also be explored. The topic is especially suitable for students ultimately interested in homeland security/treaty verification, medical physics, nuclear power plant operations, and/or radiation protection.
Research Mode: remote, hybrid (some in-person may be possible/needed for certain tasks but not required of all participants). Examples: in-person tasks could include data collection and instrument calibration.
NERS Project #6: Extended Reality Training and Virtual Reality Games for Radiation Protection
Faculty Mentor: Professor Kim Kearfott, kearfott@umich.edu
Additional Mentor: Jordan D. Noey MS, noeyjd@umich.edu
Prerequisites: Students work as part of a team; a student’s specific assignment will depend upon the background and interests of each student. Students unfamiliar with the software being used will be afforded the opportunity to learn.
Project Description: Work with a team producing a game or other extended reality experience to teach the principles of radiation protection or attract interest to the nuclear sciences. Students may become involved in the overall game or experience design, artwork, rendering of nuclear-specific objects, implementation of the realistic radiation source and detector physics, or creation of competitive aspects to the software. Unity and Blender are currently being used for the VR game, which is implemented on an Oculus Quest. Uptail and 3-D cameras are employed for the separate extended reality training experience.
Research Mode: remote. Software, 3D cameras, and virtual reality display systems will be made available to students for the performance of the project at locations of their choosing.
NERS Project #7: Design of an Intelligent Radiation Awareness Drone (iRAD)
Faculty Mentor: Professor Kim Kearfott, kearfott@umich.edu
Prerequisites: Students work as part of a team; a student’s specific assignment will depend upon the background and interests of each student. Students unfamiliar with any software being used will be afforded the opportunity to learn.
Project description: An affordable drone capable of carrying a radiation detector, and having its movements controlled based on collected information, is being designed and constructed from parts. The Intelligent Radiation Awareness Drone (iRAD) is ultimate to be provided to high schools to interest students in both nuclear sciences and robotics. Before that can occur, however, technical problems involving aircraft design (motors, layout), collision avoidance (visible camera), terrain holding (LIDAR), payload interfacing (radiation and other detectors), environmental hardening, and navigational control need to be solved. This undergraduate team is assigned to work together on those tasks.
Research Mode: in-person. This project involves the physical construction of a drone, which requires in-person work to be conducted.
NERS Project #8: Light Water Reactor Modelling with MPACT/PARCS
Faculty Mentor: Brendan Kochunas, bkochuna@umich.edu
Prerequisites: US Citizenship; experience working in Linux, basic Python programming experience (desired)
Project Description: Students will learn and apply industry-standard approaches for the modeling and analysis of LWRs. In this approach, the core is represented as homogeneous nodes and simulated with a multi-physics 3D diffusion calculation. The required nodal data is computed from the transport solution of a smaller problem with explicit geometry (i.e. a fuel lattice). Recent developments in the MPACT code have made it possible to generate this data for the fuel and reflector. In this project, the student will use MPACT and PARCS to model a LWR core (e.g. NuScale, boiling water reactors, VVER-1000). The desired outcome is to achieve a complete analysis of the functioning core model for some applications. The student should expect to gain experience in nuclear reactor analysis and tools for reactor analysis.
Research Mode: in-person, hybrid
NERS Project #9: Unstructured Mesh and Neutron Transport Software Development
Faculty Mentor: Brendan Kochunas, bkochuna@umich.edu
Prerequisites: Some programming experience (C, C++, or Fortran preferred)
Project Description:
For next-generation nuclear reactor designs and multi-physics simulations, the historic use of constructive solid geometry to represent models poses major challenges. As an alternative, some researchers have begun to use unstructured meshes akin to several commercial engineering analysis tools. A new unstructured mesh library is being developed to represent complex geometries and enable the use of advanced, automated meshing algorithms. Students will create and simulate nuclear reactor models, contribute to the development of the mesh library, and perform studies on the efficacy of meshing algorithms.
This project will be co-advised by a graduate student and has the potential for ongoing research.
Research Mode: In-person, hybrid
NERS Project #10: Deciphering United States Public Support of Nuclear Energy on Twitter with Machine Learning Natural Language Processing
Faculty Mentor: Majdi I. Radaideh, radaideh@umich.edu
Prerequisites: Python Programming and basic machine learning knowledge is a plus
Project Description: Nuclear energy is one of the leading carbon-free energy sources in the nation providing about 19% of the nation’s electricity and about 50% of the nation’s carbon-free electricity. As advanced nuclear reactors are considered a central part of the nation’s carbon-free transition, we know little about the public support for building new nuclear power plants in the nation. Social media like Twitter offers a platform that provides a massive amount of data that we can use to extract public sentiment about nuclear energy. This project will teach students how to perform web scraping to collect and mine tweets that discuss nuclear energy topics. Then, students will learn how to process and tokenize Twitter data to prepare them for machine learning. Lastly, students will apply state-of-the-art deep language models such as Google’s BERT to process the data and determine public sentiment regarding nuclear power. There are plenty of skills to learn and interesting ideas that can develop from this project.
Research Mode: Hybrid
NERS Project #11: Analysis of Reinforcement Learning Robustness for Microreactor Control
Faculty Mentor: Majdi I. Radaideh, radaideh@umich.edu
Prerequisites: Python programming and basic machine learning knowledge is a plus
Project Description: Nuclear microreactors are an innovation in nuclear power that provide modular and small-scale power reactors for electricity generation in remote areas. Due to their small size, microreactors have unconventional ways of operation and control, which drive new research questions about the autonomous control of these reactors. This project explores the potential of deep reinforcement learning algorithms in offering a robust and safe operation of microreactors. Students will focus on implementing and testing various deep reinforcement learning algorithms for a microreactor design developed in-house. The conclusions of this project can open the door for a new control paradigm based on machine learning that can complement existing approaches such as PID and model predictive control.
Research Mode: Hybrid
NERS Project #12: Development of Multiobjective Optimization Algorithms for Nuclear Reactor Design
Faculty Mentor: Majdi I. Radaideh, radaideh@umich.edu
Prerequisites: Python programming and basic mathematical optimization knowledge is a plus
Project Description: To reduce nuclear reactor design, operation, and maintenance costs, nuclear engineers often employ advanced optimization algorithms based on evolutionary & swarm intelligence to solve complex, high-dimensional, and constrained optimization problems. For that purpose, we recently developed a new open-source framework called NEORL (NeuroEvolution Optimization with Reinforcement Learning), which houses novel and state-of-the-art optimization algorithms that can be applied to different disciplines including nuclear power. In this project, students will implement and benchmark multi-objective optimization algorithms such as the genetic algorithm and particle swarm optimization on a nuclear reactor design optimization problem to reduce fuel cycle costs. The students are expected to apply NEORL and enhance it with new capabilities as well as learn various data analytics and optimization skills. The conclusions of this work can help design new nuclear reactors to be more cost competitive. In addition, depending on the progress achieved, the students can join a wonderful list of NEORL collaborators here:
https://neorl.readthedocs.io/en/latest/misc/contrib.html
Research Mode: Hybrid
NERS Project #13: Software Development in the Hammer Transport Framework
Faculty Mentor: Brian Kiedrowski, bckiedro@umich.edu
Prerequisites: Basic knowledge of C++
Project Description: The University of Michigan Computational Particle Transport team is developing software to simulate neutral particle transport through matter for general engineering applications. This project will provide students with experience working on a software development team that integrates core nuclear engineering principles. Specific tasks will be determined by the needs of the team and the interests of the student.
Research Mode: In-person, remote, or hybrid
NERS Project #14: Theory and Computation of Electron Transport in Random Media
Faculty Mentor: Brian Kiedrowski, bckiedro@umich.edu
Prerequisites: Math 216 (or equivalent experience in differential equations)
Project Description: The University of Michigan Computational Particle Transport team recently developed a new methodology for efficiently simulating the transport of electrons through matter with randomly varying media with an application to medical physics, inertial confinement fusion, atmospheric or aquatic transport, porous media (e.g., foams and certain kinds of radiation detectors), etc. The current models have shortcomings that require both mathematical and computational developments to resolve. The specific project can be tailored to the interests of the student.
Research Mode: In-person, remote, or hybrid
NERS Project #15: Application of Reduced Basis Methods to Nuclear Physics Models for Cross Sections
Faculty Mentor: Brian Kiedrowski, bckiedro@umich.edu
Prerequisites: Math 216 (or equivalent experience in differential equations), basic knowledge of C++
Project Description: The University of Michigan Computational Particle Transport team is exploring a set of reduced basis methods from the field of machine learning to more efficiently calculate model parameters for nuclear physics calculations for the generation of neutron interaction coefficients or cross sections. The student will work with the team to implement and test these models.
Research Mode: In-person, remote, or hybrid
NERS Project #16: Development of Monte Carlo Algorithms for Reactor Transient Analysis
Faculty Mentor: Brian Kiedrowski, bckiedro@umich.edu
Prerequisites: Math 216 (or equivalent experience in differential equations), basic knowledge of C++
Project Description: The University of Michigan Computational Particle Transport team is developing efficient methods to simulate time-dependent nuclear fission reactor transients with Monte Carlo calculations. While the current method is much faster than existing approaches, there is still a need to further accelerate the method. Students will develop mathematics and theory to support parallel-in-time algorithms, which is a relatively new area of study.
Research Mode: In-person, remote, or hybrid
NERS Project #17: How can we study a Coronal Mass Ejection in the Laboratory?, Carolyn Kuranz
Faculty Mentor: Carolyn Kuranz, ckuranz@umich.edu
Prerequisites:
Project Description: Interplanetary Coronal Mass Ejections (ICMEs) are solar explosions that are the source of geomagnetic storms. Learning how ICMEs interact and change throughout their time traveling through the solar system will give us a better understanding of how to predict the timing of them and protect our satellites and electricity grid from particularly powerful ones. To study ICMEs, we create a scaled experiment using the Big Red Ball (BRB) at the Wisconsin Plasma Physics Laboratory (WIPPL). Students will learn data analysis techniques and gain familiarity with writing simulations.
Research Mode: In Lab, Hybrid