Failure is Not an Option: Avoiding Operational Disruptions with Mechanistic and Data-driven Damage Prognostics (plenary session)

We are in an age where pervasive sensing, high communication bandwidth, and advances in AI have arrived at industrial equipment. The question is how one can leverage these advances for operational gain. To uphold operational functionality, these techniques flow into a Condition-Based Maintenance (CBM) strategy where maintenance is only performed on evidence of need identified through direct or indirect monitoring. Knowledge of an asset’s condition and how it will evolve is required such that the remedial action can be prescribed with sufficient lead time to minimize the cost and operational impact of the occurrence of a potential disruption. This strategy differs from “on-condition” maintenance in that an understanding of how much time is available before the asset loses functionality can be leveraged. The basic concept entails collecting and assessing data from NDE inspections and in-situ sensors to estimate remaining life of the system in question. This is done using either mechanistic, physics-based models or, as suitable, data-driven AI techniques. This talk lays out a roadmap of the tools and methods that are to be used to realize the promise of making failure not an option.