TItle: Physics-informed Machine Learning for SHM
Organisers: Alice Cicirello (University of Cambridge), Elizabeth Cross (University of Sheffield), Eleni Chatzi (ETH Zürich)
There is a need for developing robust Structural Health Monitoring (SHM) tools able to support tasks at the higher ranks of the SHM hierarchy such as timely damage detection and characterization, remaining useful life assessment, decision making and maintenance scheduling. A main research challenge to this end is overcoming poor generalisation performance and physically inconsistent or implausible predictions by developing approaches integrating physics (first principles) knowledge with Machine Learning (ML). This is of particular importance for complex and critical engineering structures, where the system “as-is” often deviates significantly from the original “as-designed” prototype, such as aerostructures, bridges, nuclear power plant and wind turbines to name a few.
This special session invites contributions on methodological techniques and industrial applications showcasing recent advances in incorporating physics in ML strategies including (but not limited to) those leveraging on observational biases (e.g. data augmentation), inductive biases (e.g. physical constraints), learning biases (e.g. inference/learning algorithm setup), and model form/discrepancy biases (e.g. equation terms describing a partially known physics-based model).