In a large-scale plant such as a nuclear power plant, many process values are measured for the purpose of monitoring its plant performance and the health of various systems. It is difficult for plant operators to constantly monitor all of the process values. We present a new data-driven method to monitor a large number of process values and to enable early detection of anomaly signs including unknown events with few false detections. In order to handle the complex plant system, we created a two-stage model composed of a time window and a deviation autoencoders which accurately predicts the normal process values. It handles a large number of process values, rapid changes of the process values such as the operation mode change, both the steady and the transient states, and the external disturbances such as exogenous noise, atmospheric temperature, etc. The time window autoencoder examines time correlations of time series data while the deviation autoencoder treats correlations of differences between the data and the former outputs. The difference between the actual process values and the predicted ones from the model, trained with normal process values, indicates the signs of abnormalities when it exceeds its threshold. We evaluated this model with simulation data of a nuclear power plant, 1.1 GW Boiling Water Reactor having 3,100 analog process values. The two-stage model showed a good anomaly detection performance with few false detections in both the steady and the transient states under the external disturbance.


A New Data Driven Method for Monitoring a Large Number of Process Values and Detecting Anomaly Signs With a Two-Stage Model Composed of a Time Window Autoencoder and a Deviation Autoencoder

Division / Sponsors: 
Power Division & Nuclear Engineering Division (NED)

Power Conference
Paper Proceedings: https://asmedigitalcollection.asme.org/POWER/POWER2020/volume/83747

International Conference on Nuclear Engineering
Paper Proceedings: https://asmedigitalcollection.asme.org/ICONE/ICONE2020/volume/83761

POWER-ICONE - Virtual Conference

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.