Building Better Monitoring Systems

In the final part of our interview with Hunter Andrews, an R&D Staff Scientist at Oak Ridge National Laboratory, he discusses how other analytical approaches, such as electrochemistry, mass spectrometry (MS), and chemometrics, are being integrated to build better monitoring systems (1).

In the below video segment, Andrews explains how his wide-ranging expertise across electrochemistry, neutron imaging, molten salts, MS, chemometrics, and machine learning (ML) helps him approach analytical challenges through “lessons learned” from multiple disciplines. He discusses how he applies an engineering mindset to his work, which involves defining problems, constraining variables, optimizing methods, and then testing solutions under real-world conditions.

Using the analogy of spectroscopy, Andrews compares each analytical technique to a color: a single method may yield a limited picture, but combining techniques produces a fuller, more accurate view. Integrating diverse tools is difficult because of differing instrument behaviors, which is why ML, chemometrics, and sensor fusion play key roles. These data-driven approaches help merge parallel signals and manage variability. The greatest challenge, Andrews noted to Spectroscopy, is ensuring ML models remain robust when faced with unexpected conditions, since real-world systems rarely behave like training data.

This video clip is the final part of our conversation with Andrews. To stay up to date on our coverage of the Winter Conference on Plasma Spectrochemistry, click here.

References

1. Oak Ridge National Laboratory, Hunter B. Andrews. ORNL.gov. Available at: https://www.ornl.gov/staff-profile/hunter-b-andrews (accessed 2026-01-06).
2. IASA, Winter Conference on Plasma Spectrochemistry. IASA. Available at: https://iasa.world/winter-plasma-conference (accessed 2026-01-06).