MACHINE LEARNING · PREDICTIVE MAINTENANCE
Jet Engine Hospital
An end-to-end predictive-maintenance system for NASA C-MAPSS turbofan engines.
The problem
Predict remaining useful life, identify near-term failure risk, detect anomalous behavior, quantify uncertainty, and turn those signals into an interpretable maintenance action.
Sensor Data
Feature Engineering
RUL Regression
Risk Classification
Anomaly Detection
Conformal Interval
Decision Policy
Dashboard
Unified dashboard
Three operating scenarios
FD001 provides the foundation, FD003 introduces multiple fault modes, and FD004 adds multiple operating conditions and multiple faults. The final interface presents a unified workflow across the three datasets.