The United States government has made investments in research to more accurately identify the bad actors’ destructive activities as they grow more skilled in their attempts to get beyond international nuclear nonproliferation protections. Organizations like the International Atomic Energy Agency (IAEA) use meticulous monitoring measures to ensure that nuclear materials covered by agreements aren’t used to develop nuclear weapons in order to address nonproliferation threats.
They also use advanced forensic techniques to identify the source of radioactive materials that law enforcement has found. These methods, however, are frequently time and labor-intensive. In order to detect threats and conduct a forensic investigation in the nuclear sector more quickly and efficiently, a new study from the Pacific Northwest National Laboratory (PNNL) utilizes machine learning, artificial reasoning, and data analytics.
PNNL conducts research, converts the research into usable real-world solutions, and attempts to come up with creative methods by integrating these computer tools with experience in nonproliferation and safeguards. PNNL is in a unique position to combine machine learning and data analytics with nuclear expertise. The IAEA is fully aware of the information the organization seeks out in order to reveal actors’ potentially harmful conduct.
Nuclear proliferation must be stopped. Work is required, including audits on nuclear materials and inquiries into those handling nuclear items. Techniques based on data analytics can be used to facilitate this. The Department of Defense, the Mathematics for Artificial Reasoning in Science (MARS) Initiative, and the National Nuclear Security Administration (NNSA) have all provided funding for PNNL to work on a number of projects that would improve nuclear nonproliferation and safeguards.
Identifying nuclear material diversion
Used nuclear fuel is taken to facilities for nuclear reprocessing, where it is separated into waste products. The chemicals that can be recycled as fresh nuclear reactor fuel are then created using the products. These substances contain uranium and plutonium, both of which are components of nuclear bombs. To ensure that no nuclear materials are being used to build nuclear weapons, the IAEA keeps an eye on nuclear facilities. This often includes both routine inspections and the collecting of samples for later destructive assays.
There was a collaboration with Sandia National Laboratories to create a virtual representation of a reprocessing facility. A machine learning model was also created to recognize patterns in process data indicative of nuclear material diversion. The model’s performance in this simulated setting was encouraging. Although it is improbable that this strategy would be utilized in the near future, this technique offers a promising beginning that will complement current safety measures.
Analyzing images to identify the source of nuclear materials
Law enforcement may occasionally come upon radioactive material that is unregulated and of unknown provenance, whether they are working in the United States or anywhere else. Finding out the origins of the material in these situations is vital since the stuff retrieved might only represent a percentage of the material that is in danger of being trafficked. To increase nuclear security in the event of such instances, the IAEA keeps a database and encourages nations to work together to combat illegal nuclear trafficking.
One analytical technique utilized in this crucial effort is the forensic study of radioactive materials. It might take some time for organizations to incorporate machine learning techniques into their procedure for identifying nuclear threats, as it is obvious that these technologies can influence and streamline the procedure.