Seismic and acoustic data are of value in detecting movements of the Earth’s crust during earthquakes and to evaluate the condition of bridges and other infrastructure. Such data can also find application in nuclear nonproliferation monitoring, as demonstrated by researchers from Expedition Technology Inc. and U.S. Oak Ridge National Laboratory (ORNL).

Sensors placed 50 m distant from the High Flux Isotope Reactor at ORNL collected data in a minimally invasive manner for a period of one year. The resulting data provided a record of the mechanical signatures of vibrating equipment such as fans and pumps at the cooling tower. The measurements were sufficientlySensors were placed 50 m from the High Flux Isotope Reactor. Source: ORNLSensors were placed 50 m from the High Flux Isotope Reactor. Source: ORNL accurate to determine the operational status of the facility and to verify that the nuclear system is operating as declared.

Machine learning models were applied to the data to classify the reactor’s operational state and power level. Inclusion of combined seismic and acoustic data was shown to determine whether the reactor was on or off with 98% accuracy and to gauge transition states and power levels with 66% accuracy.

The researchers note in Seismological Research Letters that both seismic and acoustic data could be integrated with other observations to improve monitoring performance.

To contact the author of this article, email