With seismic and acoustic data recorded by remote sensors near ORNL’s High Flux Isotope Reactor, researchers could predict whether the reactor was on or off with 98% accuracy. Credit: Nathan Armistead/ORNL, U.S. Dept. of Energy
An Oak Ridge National Laboratory team developed a novel technique using sensors to monitor seismic and acoustic activity and machine learning to differentiate operational activities at facilities from “noise” in the recorded data.
Using the lab’s High Flux Isotope Reactor as a testbed, the researchers placed remote sensors near the facility and continuously recorded data. Their published results showed they could predict whether the reactor was on or off with 98% accuracy. They could also tell whether seismo-acoustic activity was coming from reactor-specific operations or other sources, such as equipment vibrations from a nearby cooling tower.
“We got creative with the tools, and we were able to tease out the information from that seismic noise — a technique that worked well,” ORNL’s Monica Maceira said.
The team’s new approach could be used as a protective measure for sensitive facilities and nonproliferation applications and as a monitoring tool for building structural health.
