Technology Category
Data Mining & Knowledge Management: Anomaly/Outlier DetectionMachine Learning: Time-Series/Data StreamsIntelligent Robots: Multimodal Perception & Sensor Fusion
Explainable Unsupervised Multi-Anomaly Detection and Temporal Localization in Nuclear Times Series Data with a Dual Attention-Based Autoencoder
🤖 AI Summary
Nuclear power plants employing novel reactor designs generate massive multivariate time-series data, yet suffer from severe scarcity of anomalous samples and authentic labeled data; moreover, existing black-box anomaly detection models lack interpretability, hindering their deployment in safety-critical applications. To address these challenges, we propose an unsupervised anomaly detection framework built upon an LSTM-based autoencoder, integrated with dual-dimensional attention mechanisms—feature-wise and temporal—that jointly identify anomalous sensors and localize anomalous time intervals without supervision. Feature attention weights critical monitoring variables, while temporal attention pinpoints discriminative time segments, collectively enhancing diagnostic transparency. Evaluated on real-world data from the PUR-1 research reactor, our method achieves high detection accuracy across diverse complex anomalies and provides clear spatiotemporal attribution, thereby enabling trustworthy autonomous monitoring for compact reactor systems.
Technology Category
Application Category
📝 Abstract
The nuclear industry is advancing toward more new reactor designs, with next-generation reactors expected to be smaller in scale and power output. These systems have the potential to produce large volumes of information in the form of multivariate time-series data, which could be used for enhanced real-time monitoring and control. In this context, the development of remote autonomous or semi-autonomous control systems for reactor operation has gained significant interest. A critical first step toward such systems is an accurate diagnostics module capable of detecting and localizing anomalies within the reactor system. Recent studies have proposed various ML and DL approaches for anomaly detection in the nuclear domain. Despite promising results, key challenges remain, including limited to no explainability, lack of access to real-world data, and scarcity of abnormal events, which impedes benchmarking and characterization. Most existing studies treat these methods as black boxes, while recent work highlights the need for greater interpretability of ML/DL outputs in safety-critical domains. Here, we propose an unsupervised methodology based on an LSTM autoencoder with a dual attention mechanism for characterization of abnormal events in a real-world reactor radiation area monitoring system. The framework includes not only detection but also localization of the event and was evaluated using real-world datasets of increasing complexity from the PUR-1 research reactor. The attention mechanisms operate in both the feature and temporal dimensions, where the feature attention assigns weights to radiation sensors exhibiting abnormal patterns, while time attention highlights the specific timesteps where irregularities occur, thus enabling localization. By combining the results, the framework can identify both the affected sensors and the duration of each anomaly within a single unified network.
Problem
Research questions and friction points this paper is trying to address.
Detecting anomalies in nuclear reactor time-series data
Localizing abnormal events temporally and across sensors
Providing explainable unsupervised learning for safety-critical systems
Innovation
Methods, ideas, or system contributions that make the work stand out.
Dual attention mechanism for anomaly detection
Unsupervised LSTM autoencoder framework
Feature and temporal localization capabilities
🔎 Similar Papers
2024-09-03TempXAI@PKDD/ECMLCitations: 0
