How gayathriravi-codes/uav-engine-digital-twin works
AeroTwin is a UAV-engine digital twin for telemetry-based health monitoring, fault diagnosis, remaining-useful-life estimation, and predictive maintenance. Its principal workflow loads a flight in the dashboard, sends telemetry through integrated inference, and presents engine health and operator guidance; mission recommendations and counterfactual what-if estimates extend that workflow. Simulation and physics-based residual analysis provide distinct supporting capabilities. The sampled code supports the main inference and what-if links; unsampled dashboard behavior and frontend-to-backend wiring are not assumed.
Main entry points: app.py Monitoring dashboard · engine_inference.py Integrated inference · whatif_engine.py What-if analysis · train_rul_ensemble.py RUL ensemble
Components 15
Operator experience
- Monitoring dashboard
dashboard/app.py - 3D engine viewer
aerotwin-frontend/src/App.jsx
Engine inference
- Integrated inference
models/classification/engine_inference.py - Fault diagnosis
models/classification/predict.py - RUL ensemble
models/rul/train_rul_ensemble.py - Recovery detection
models/classification/recalibration_speed.py - Sensor trust
sensor_trust.py - RUL prediction API
models/rul/app.py
Maintenance decisions
- Mission recommendation
mission_reliability.py - What-if analysis
models/rul/whatif_engine.py
Simulation and physics
- Physics residuals
models/physics_residual.py - Healthy flight simulation
simulator/simulate_healthy.py - Fault injection
simulator/fault_injectors.py
Other components
- Operator
- Flight telemetry
Connections 14
- Operator → Monitoring dashboard: selects flight
- Flight telemetry → Monitoring dashboard: loads telemetry
- Monitoring dashboard → Integrated inference: requests inference
- Integrated inference → Fault diagnosis: classifies fault
- Integrated inference → RUL ensemble: estimates RUL
- Integrated inference → Recovery detection: checks recovery
- Monitoring dashboard → Sensor trust: checks sensors
- Monitoring dashboard → Physics residuals: computes residuals
- Monitoring dashboard → Mission recommendation: requests action
- Monitoring dashboard → What-if analysis: requests scenarios
- What-if analysis → Healthy flight simulation: simulates flight
- What-if analysis → Fault injection: injects fault
- What-if analysis → RUL ensemble: predicts RUL
- Monitoring dashboard → Operator: shows assessment