In this webinar, we share results from the DynFlex project, where we developed AI-based methods for condition monitoring and early fault detection in PtX facilities. Without heavy maths, we’ll explain how a deep-learning model can learn what “normal operation” looks like and flag deviations as they emerge.
We’ll show how we tested the approach on simulated chemical-reactor processes, including a methanol reactor digital twin: how the model maps a plant’s operating modes, how different faults show up, and how quickly they can be detected.
Finally, we’ll look ahead: beyond raising alarms, the same model could give operators an intuitive “map” of the plant’s state and help explain why something looks wrong, a step towards more transparent predictive maintenance.
The webinar is primarily aimed at Industry professionals interested in AI-based condition monitoring. No machine-learning background required.
Étienne Bourbeau
AI Specialist, PhD
Alexandra Institute
+45 93 52 16 77
etienne.bourbeau@alexandra.dk
This project has received funding from MissionGreenFuels through Innovation Fund Denmark
Måske kunne disse events også have din interesse:
6. NOVEMBER 2026 · ONLINE
Som rådgiver skaber du værdi for dine kunder, når du gør dem i stand til at løfte cybersikkerheden i hele organisationen. Denne masterclass klæder dig på til opgaven.
10. NOVEMBER 2026 · KØBENHAVN
Agentic AI and foundation models are creating new possibilities for how we work with bioimaging, biodata and preclinical research.
11. NOVEMBER 2026 · SLAGELSE
Bliv klogere på, hvordan du styrker cybersikkerheden i din produktionsvirksomhed. DAMRC, Slagelse Erhverv og Bright Bird inviterer til workshop med fokus på at identificere risici og omsætte dem til konkrete handlinger, der styrker din virksomheds sikkerhed.