In a recent article published in the journal Nuclear Engineering & Technology, researchers studied the performance of a nuclear facility monitoring system using a multi-sensor network and artificial intelligence algorithm. The research aims to enhance the monitoring of radioactive materials within nuclear facilities, improving safety and security.
Study: Advanced Monitoring System for Nuclear Facilities. Image Credit: Peteri/Shutterstock.com
Background
The increasing use of nuclear and radiation technologies has underscored the critical importance of radiation safety and monitoring in nuclear facilities. With the potential risks associated with radioactive materials, the need for robust monitoring systems that can rapidly respond to radiation incidents is paramount. Traditional monitoring methods may not always provide the level of accuracy and efficiency required to ensure the safety and security of nuclear facilities.
In this context, the development of a sophisticated nuclear facility monitoring system (NFMS) becomes essential to address the challenges posed by the storage and handling of radioactive materials. The NFMS aims to enhance the detection, tracking, and identification of radioactive sources within nuclear facilities, thereby improving overall safety protocols and emergency response capabilities.
The Current Study
The nuclear facility monitoring system evaluated in this study was designed with a multi-sensor network and artificial intelligence algorithm to enhance the tracking and identification of radioactive materials within nuclear facilities. The monitoring system's architecture consisted of multiple detectors strategically placed based on the facility's geometry.
The multi-sensor network utilized NaI(Tl) detectors and a Field-Programmable Gate Array (FPGA)-based Data Acquisition (DAQ) system. The detectors were arranged in a 3 × 2 array configuration to monitor a 5 × 5 drum setup within the facility. Each detector was responsible for monitoring a specific area, with overlapping monitoring regions between adjacent detectors to ensure comprehensive coverage.
To optimize the performance of the multi-sensor network, the Geant4 application for tomographic emission (GATE) toolkit was employed. This tool allowed for the simulation and analysis of radiation interactions within the monitoring system, aiding in the efficient placement of detectors for maximum coverage and sensitivity.
The evaluation of the system's energy resolution and sensitivity was conducted to ensure accurate tracking of radioactive materials and identification of nuclides. Experimental tests were performed in a test storage facility to assess the localization accuracy of the NFMS. The measured counts from each detector were analyzed using an artificial neural network (ANN) algorithm to determine the localization accuracy of radioactive sources.
The AI-based ANN algorithm played a crucial role in processing the measured counts and identifying changes in radioactivity levels. By analyzing the ratios of measured counts from different detectors, the algorithm could detect variations in radioactivity and accurately track the location of radioactive materials within the facility.
Results and Discussion
The performance evaluation of the nuclear facility monitoring system using the multi-sensor network and artificial intelligence algorithm yielded promising results. The energy resolution and sensitivity of the NaI(Tl) detectors were measured to assess the system's capability to track radioactive materials accurately. The experiment demonstrated energy resolutions of 16.41 %, 8.43 %, 8.39 %, and 4.61 % for Co-57, Na-22, Cs-137, and Co-60 nuclides, respectively. These results confirmed the system's ability to identify nuclides within the specified energy range.
Normalization procedures were conducted to determine the sensitivity of individual detectors within the multi-sensor network. By normalizing the counts and analyzing the ratio of counts from each detector, the system could effectively track changes in radioactivity levels. Experimental tests at different locations revealed a ratio of measured counts based on the distance between the detector and the radiation source, providing valuable insights into the system's performance under varying conditions.
The accuracy of the ANN algorithm for location tracking was evaluated through experiments monitoring a 3 × 2 array of drums within a scaled-down facility. The results indicated that the ANN algorithm achieved an accuracy of over 99 % in tracking the location of radioactive sources. This high level of accuracy is essential for ensuring the system's reliability in detecting and monitoring radioactive materials within nuclear facilities.
The optimization of the multi-sensor network configuration, as demonstrated through the 3 × 2 array of detectors, proved to be effective in monitoring a 5 × 5 drum setup. The overlapping monitoring areas between adjacent detectors enhanced the system's coverage and sensitivity, enabling comprehensive tracking of radioactivity changes. The integration of the Geant4 application for tomographic emission (GATE) toolkit facilitated the efficient placement of detectors, further enhancing the system's performance.
Conclusion
The research demonstrates the successful performance evaluation of the nuclear facility monitoring system, highlighting the benefits of using a multi-sensor network and AI algorithm. The system that was developed by the research team shows promise when it comes to improving the accuracy and efficiency of monitoring radioactive materials within nuclear facilities, contributing to enhanced safety and security measures.
Journal Reference
Min Kyu Baek, Insoo Kang, et al. (2024). Performance evaluation of a nuclear facility monitoring system using multi-sensor network and artificial intelligence algorithm. Nuclear Engineering and Technology, ISSN 1738-5733. https://doi.org/10.1016/j.net.2024.06.010, https://www.sciencedirect.com/science/article/pii/S1738573324002699