Online Monitoring System to Support Autonomous Remote Microreactor Operations

Microreactors are factoryfabricated and transportable reactors that can be used for electricity production and thermal energy generation in remote offgrid locations. They are designed to be selfregulating autonomous systems that do not require a large number of onsite staff. To realize the full benefits of autonomous operation, microreactors need online monitoring systems that provide continuous performance verification of critical components at a level of automation that is not available today in the nuclear industry. In addition, microreactor process conditions are harsher relative to existing light water reactors, and existing nuclear sensors may not be capable of withstanding the longterm effects of these stressors in a microreactor. To avoid frequent sensor replacement as a result of calibration drift or premature degradation, durable alternative sensor technologies must be identified and evaluated for microreactor applications. To facilitate the safe and efficient deployment of microreactors, a research and development effort is proposed involving the testing of sensors and embedded instruments that could be installed within microreactors and used in conjunction with an online monitoring system that verifies the health of critical components. In support of this system, artificial intelligence, machine learning, and other advanced analysis techniques will be identified and developed specifically for microreactor applications. This work will result in an online monitoring system to support remote autonomous microreactor operations. The Phase I work successfully demonstrated the feasibility of employing artificial intelligence and machine learning techniques to provide identification and diagnosis of sensor and process anomalies using data from a variety of sources. In addition, the Phase I work highlighted the challenges associated with existing instrumentation to provide embedded sensor measurements and structural health monitoring. The research and development effort proposed herein will employ a handson approach to include demonstration of artificial intelligence and machine learning techniques to identify, diagnose, and predict sensor and/or process anomalies, experimental laboratory work to characterize the static and dynamic performance of process and nonprocess sensors that may be used in microreactors, and development of an online monitoring system that can provide realtime condition assessment of microreactors and inform decisionmaking of the autonomous control system. The results of this project will support the development and deployment of microreactors in the United States and the export of American microreactor technologies to other countries. The results will be broadly applicable to microreactors, advanced reactors, small modular reactors, and the existing nuclear fleet for improved insitu sensor performance verification as well as process and component health monitoring, diagnostics, and prognostics.