Predictive maintenance (PdM) is one of the strategies that has shown great potentials in achieving substantial cost savings and enhancing the economic competitiveness of nuclear power plants in the current energy market. PdM strategy taking advantage of advancements in machine learning (ML) technologies have demonstrated ability in handling high dimensional and multivariate data and in extracting hidden relationships within data in industrial environments. While ML technologies show great potentials, their lack of explainability, especially in considering multiple aspects in human-scale explanation-giving tasks, is one of the major hurdles to their adoptions. The research presented in this paper develops an explainable ML solutions by accounting for four attributes of explainable artificial intelligence, including the contextual factors, explainable model options, post-hoc explanations for black-box models using Shapley additive explanations and local interpretable model-agnostic explanations, and graphical user interface for human cognitive capacity and limitations. This tool is then applied to the conducting of PdM tasks for a circulating water system in a nuclear power plant.

As part of an initiative to steward research, development, and innovation into the nuclear fuel cycle, Idaho National Laboratory is building the Beartooth test bed. Beartooth will include a cascade of centrifugal contactors, glove box lines, and solidification and dissolution equipment to aid in the progression of novel separation techniques and to provide hands-on opportunities to early-career separation engineers. Beartooth will incorporate novel monitoring techniques using sensors and machine learning algorithms to inform process operators of separation conditions. This research is examining unconventional monitoring technologies such as incorporating acoustic microphones, vibration sensors, infrared cameras, red-green-blue color sensors, among others into nuclear separation processes. Research is ongoing in the use of machine learning methods to detect faults and alert operators of typical and anomalous events. Results from this research have the potential to impact safeguards-by-design efforts and real-time decision making. This overview will detail preliminary measurement results from acoustic, vibration, and color sensors.

As part of an initiative to steward research, development, and innovation into the nuclear fuel cycle, Idaho National Laboratory is building the Beartooth test bed. Beartooth will include a cascade of centrifugal contactors, glove box lines, and solidification and dissolution equipment to aid in the progression of novel separation techniques and to provide hands-on opportunities to early-career separation engineers. Beartooth will incorporate novel monitoring techniques using sensors and machine learning algorithms to inform process operators of separation conditions. This research is examining unconventional monitoring technologies such as incorporating acoustic microphones, vibration sensors, infrared cameras, red-green-blue color sensors, among others into nuclear separation processes. Research is ongoing in the use of machine learning methods to detect faults and alert operators of typical and anomalous events. Results from this research have the potential to impact safeguards-by-design efforts and real-time decision making. This overview will detail preliminary measurement results from acoustic, vibration, and color sensors.

A tremendous commitment of resources is needed to acquire, understand and apply battery data in terms of performance and aging behavior. There are many state of performance (SOP) and state of health (SOH) metrics that are useful to guide alignment of batteries to end-use, yet how these metrics are measured or extracted can make the difference between usable, valuable datasets versus data that lacks the necessary integrity to meet baseline confidence levels for SOP/SOH quantification. This work will speak to 1) types of data that support SOP and SOH evaluations on mechanistic terms, 2) measurement conditions needed to assure high data integrity, 3) equipment limitations that can compromise data high fidelity, and 4) the impact of cell polarization on data quality. A common goal in battery research and field use is to work from a data platform that supports economical paths of data capture while minimizing down-time for battery diagnostics. An ideal situation would be to utilize data obtained during normal daily use (“pulses or cycles of convenience”) without stopping the daily duty cycles to perform dedicated SOP/SOH diagnostic routines. However, difficulties arise in trying to make use of daily duty cycle data (denoted as cycle-by-cycle, CBC) that underscores the need for standardization of conditions: temperature and duty cycles can vary over the course of a day and throughout a week, month and year; polarization can develop within an immediate cycle and throughout successive cycles as a hysteresis. If CBC data is envisioned as a data source to determine performance and aging trends, it should be recognized that polarization is a frequent consequence of CBC and thus makes it difficult to separate reversible and irreversible components to metrics such as capacity loss and resistance increase over aging. Since CBC conditions can have a major impact on data usability, we will devote part of this paper to CBC data conditioning and management. Differential analyses will also be discussed as a means to detect changing trends in data quality. Our target cell chemistries will be lithium-ion types NMC/graphite and LMO/LTO.

Light-water reactor operations and maintenance (O&M) costs are prohibitively high, thus contributing to the premature decommissioning of nuclear power plants (NPPs). This is partly due to how the equipment is monitored. In recent years, cloud computing has emerged as a dominant technology by virtue of its low costs, computing and storage adaptability, and ability to host applications over numerous types of virtual infrastructures. Cloud computing can be a cost-effective alternative to onsite storage and diagnostics. This paper conducts a techno-economic assessment of a provisional cloud deployment architecture for a NPP predictive monitoring (PdM) system. The cloud-based monitoring system would enable maintenance and diagnostics (M&D) analysts and other authorized plant users to remotely monitor equipment functionality so as to enable PdM practices and early detection of faults. The Microsoft Azure cloud platform is included in the proposed cloud architecture to provide data processing and storage, sensor device networking, and database management; however, this analysis could be extended to other cloud computing service providers as well. For the techno-economic assessment, technical feasibility is measured in terms of network performance metrics such as response time, latency, and throughput, whereas economic feasibility is measured in terms of operational costs and capital expenditures. Finally, this report covers certain regulatory and security aspects that may concern licensees looking to implement cloud computing. The report focuses on the integration of sensor database storage, the application of cloud resources to PdM, and the identification of technological and economic hurdles associated with moving to a cloud-computing-based architecture.

The domestic nuclear power plant fleet has relied on labor-intensive and time-consuming preventive maintenance programs, thus driving up operation and maintenance costs to achieve high-capacity factors. Artificial intelligence and machine learning can help simplify complex problems, such as diagnosing equipment degradation, to enable more effective decision-making. Benefits will be felt not only within existing analog and digital instrumentation and control, but also work processes, the integration of people with technology, and most importantly, the business case. Together, these hold promise to make nuclear power more efficient and reduce costs associated with operation and maintenance. While the artificial intelligence and machine learning technologies hold significant promise in the nuclear industry, there are challenges or barriers to their adoption. This report outlines the those different machine learning adoption barriers (categorized as historical, technical, economic, regulatory, and user) that the industry must overcome to realize the full benefits of artificial intelligence and machine learning capabilities for long-term economic sustainability. This report also provides solutions for some of these barriers by focusing on improving the explainability of machine learning to encourage trust from the end-user. Trust and explainability are essential to machine learning adoption. This report focuses on research-developed solutions to some of these barriers while analyzing a non-safety-related system, namely the circulating water system. This system frequently experiences waterbox fouling which our models preemptively diagnoses then explains to the operator how those conclusions were reached. This report presents and discusses the inherent trade-off between machine learning performance (in terms of accuracy) and explainability, where highly accurate machine learning methods (such as deep-learning) are the least explainable, and the most explainable methods (such as decision trees) are the least accurate. In addition, explainability of artificial intelligence techniques in terms of transparency and post-hoc metrics are discussed. This report outlines the importance of data novelty and value of new information in evaluating both the explainability and trustworthiness. Novelty detection helps to establish consistency or inconsistency of the new data with respect to the training data. On the other hand, value of information could be a part of the user-centric visualization recommendation system that request additional information to be collected, thereby strengthening the machine learning outcomes. During this project, a copyrighted user-centric visualization that aligns with a human-in-the-loop approach was developed. The user-centric visualization presents different levels of information and can be tailored as per user credentials to gain user confidence. One of the salient features of the user-centric visualization is it presents machine learning methods with explainability metrics. A simplified version of the user-centric visualization was presented to 32 users with varying levels of machine learning expertise. Feedback was solicited to test the hypothesis that the app contained sufficient explainability and that the users would trust the algorithm. Overall, the app was positively received, and the hypothesis was supported. This report discusses the trust-but-verify framework – a potential approach to build user trust artificial intelligence. The framework discusses trust from the human level to artificial intelligence level. The fundamental premise of the trust but verify framework is derived from an observation of nuclear safety culture (i.e., nuclear power plant personnel do not rely on a singular source of data to make a decision). This also ties back to the user-centric visualization that presents different levels of information to achieve both explainability and trustworthiness of artificial intelligence. Even so, the adoption of artificial intelligence and machine learning in the nuclear industry faces additional barriers, namely regulatory and stakeholder readiness. To overcome these challenges, new solutions must gain regulatory approval and cater to stakeholder needs. The Nuclear Regulatory Committee has a 5-year strategic plan which prepares them for reviewing artificial intelligence technologies in licensee submissions. Early and frequent engagement with the regulator is encouraged. Additionally, artificial intelligence solutions should incorporate human-in-the-loop considerations and offer explainability. Stakeholders must prepare by hiring or training staff to adapt to advancing technology in everyday plant tasks.

Nuclear power plants collect and store large volumes of heterogeneous data from various components and systems. With recent advances in machine learning (ML) techniques, these data can be leveraged to develop diagnostic and short-term forecasting models to better predict future equipment condition. Maintenance operations can then be planned in advance whenever degraded performance is predicted, thus resulting in fewer unplanned outages and the optimization of maintenance activities. This enables lower maintenance costs and improves the overall economics of nuclear power. This paper focuses on developing a short-term forecasting process that leverages a feature selection process to distill large volumes of heterogeneous data and predict specific equipment parameters. A variety of feature selection methods, including Shapley Additive Explanations (SHAP) and variance inflation factor (VIF), were used to select the optimal features as inputs for three ML methods: long short-term memory (LSTM) networks, support vector regression (SVR), and random forest (RF). Each combination of model and input features was used to predict a pump bearing temperature both 1 and 24 hours in advance, based on actual plant system data. The optimal inputs for the LSTM and SVR were selected using the SHAP values, while the optimal input for the RF consisted solely of the response variable itself. Each model produced similar 1-hour-ahead predictions, with root mean square errors (RMSEs) of roughly 0.006. For the 24-hour-ahead predictions, differences could be seen between LSTM, SVR, and RF, as reflected by model performances of 0.036 +- 0.014, 0.0026 +- 0, and 0.063 +- 0.004 RMSE, respectively. As big data and continuous online monitoring become more widely available, the proposed feature selection process can be used for many applications beyond the prediction of process parameters within nuclear infrastructure.