How garudra/aeris---piston-engine-digital-twin-sih-2026-sih26054 works
AERIS is a piston-engine digital twin with live telemetry, fault injection, ML health prediction, and a React cockpit. The main workflow is cockpit controls to the backend API, which drives the physics simulator, streams telemetry, runs inference, and persists alert records; the cockpit presents live readings, trends, engine visualization, and fault history. Dataset generation and ML training are separate supporting capabilities. Connections follow the excerpts and README; training/data wiring is not asserted where source evidence was unavailable.
Main entry points: App.tsx Cockpit dashboard · api.py API and stream · train_pipeline.py Model training · generate_dataset.py Dataset generator
Components 14
Live backend
- API and stream
Backend/api.py - Engine simulator
Backend/engine_model.py
ML and data
- ML inference
Backend/ml/predict.py - Model training
ML/train_pipeline.py - Dataset generator
Backend/generate_dataset.py - Engine dataset
Database/engine_dataset.csv - Fault catalog
Database/fault_catalog.json
Cockpit UI
- Cockpit dashboard
Frontend/src/App.tsx - Telemetry and controls
Frontend/src/dashboard/engine.ts - Engine visualization
Frontend/src/dashboard/EngineSVG.tsx - Telemetry charts
Frontend/src/dashboard/charts.tsx - UI entry point
Frontend/src/main.tsx
Alert history
- Fault history store
Backend/db.py
Other components
- Operator
Connections 13
- Operator → Cockpit dashboard: operates
README.md - UI entry point → Cockpit dashboard: renders
Frontend/src/main.tsx - Cockpit dashboard → Telemetry and controls: uses controls
Frontend/src/App.tsx - Cockpit dashboard → Engine visualization: renders
Frontend/src/App.tsx - Cockpit dashboard → Telemetry charts: renders
Frontend/src/App.tsx - Telemetry and controls → API and stream: sends controls
Frontend/src/dashboard/engine.ts - API and stream → Engine simulator: drives simulation
Backend/api.py - API and stream → ML inference: predicts health
Backend/api.py - API and stream → Fault history store: stores alerts
Backend/api.py - Dataset generator → Engine simulator: simulates runs
Backend/generate_dataset.py - Dataset generator → Engine dataset: writes dataset
README.md - Model training → Engine dataset: trains from data
- Model training → ML inference: exports models
README.md