Friday, 11 December 2026
12:45 | Join in |
13:00 | Welcome, Motivation & Introduction to Predictive Maintenance |
13:15 | Understanding degradation in multivariate time-series data |
– introduction to the NASA C-MAPSS datase | |
– run-to-failure trajectories, sensors and operating conditions | |
– identifying degradation and anomalous behaviour | |
Hands-on lab: exploring engine trajectories and identifying useful sensor signals | |
14:15 | Break |
14:30 | From degradation detection to Remaining Useful Life prediction |
– constructing an RUL target | |
– feature/window-based time-series modelling | |
– evaluating RUL predictions and understanding model errors | |
Hands-on lab: training a lightweight RUL prediction model on C-MAPSS | |
15:30 | Break |
15:45 | Explaining predictive-maintenance decisions |
– explaining which sensors and time-series characteristics drive an RUL prediction | |
– applying SHAP / feature-attribution techniques to selected engine cases | |
– comparing healthy, degrading and near-failure examples | |
Collaborative discussion: what makes an explanation useful for a maintenance engineer? | |
16:30 | End of course |