Fault Detection Methods and Early Data in New UK Cosmic Ray Neutron Monitor

In recent years, the impacts of solar activity on the earth’s surface, atmospheric and orbital environments have become increasingly important. Governments around the world have identified “space weather” as a potentially significant risk to critical infrastructure. However, the monitoring and forecasting capabilities for these events are currently rudimentary compared to terrestrial weather.

A global network of galactic and solar cosmic ray ground level monitors produce data for academic study and Ground Level Enhancement (GLE) event alerting. A GLE is typically characterised by a sudden large increase flux of fast neutrons over a wide area of the Earth’s surface for a period of 15 minutes or longer followed by a relatively gradual drop to the quiescent level. The flux of energetic subatomic particles during these events can degrade solar arrays, damage electronic components or cause single event effects (malfunctions) in semiconductor devices, potentially leading to significant disruption.

However, this network is built primarily using a neutron monitor (NM) design from 1964, named the NM-64. In the intervening years computer driven digital acquisition and processing have been added, but the monitor itself remains the same. Consequently, a new cosmic ray NM has been commissioned, the design for which has been optimised using Monte Carlo N-Particle simulations and experimentally validated. The counting performance of this design matches the NM-64 whilst using alternative non-toxic detectors, and being significantly smaller and lighter.

The new NM-2023, soon to be installed at a UK Meteorological Office site, will send its data to the Met Office Space Weather Operations Centre (MOSWOC) and the Neutron Monitor DataBase (NMDB) for global dissemination. Supporting real-time data processing and transmission to recipients facilitates their integration into operation forecast products, such as the nowcasting of atmospheric radiation exposure for end-users in the aviation sector.

As the accuracy of NM-2023 data is vital, a series of algorithms are applied to perform error checking and correction, along with air pressure correction and data formatting, before the data is disseminated. Two anomaly detection methods are employed, one on the immediate data and the other on the long-term trends. Multiple approaches have been developed for the immediate analysis, one using Principal Component Analysis. Weekly averages of count rates and measurements of neutron multiplicity form some of the long-term analysis. Concurrently with code development, a prototype for the NM-2023 has been gathering cosmic ray data at multiple locations in the UK. The results track trends seen by nearby NM-64s. The presentation will cover a discussion of the anomaly correction algorithms, in combination with a comparison of some data from the current NM-64 network with both the prototype and early results from the NM-2023.