How tjmhmdaarif/windsense-ai works
WindSense AI is a Streamlit wind-farm monitoring and alarm-analysis application. Operators enter through authentication and reach a realtime dashboard, where alarm classification and anomaly analysis support prioritization, root-cause investigation, and production/downtime analytics. The README documents these dashboard capabilities, but sampled source does not establish all their internal wiring; the graph therefore keeps unsampled dashboard functions under the page rather than inventing a processing pipeline.
Main entry points: 1_Realtime.py Realtime dashboard · app.py Streamlit entry · login.py Login and signup · anomaly_detector.py Anomaly detection
Components 20
Application flow
- Streamlit entry
app.py - Login and signup
pages/login.py - Realtime dashboard
pages/1_Realtime.py - OPC UA simulator
utils/opcua_simulator.py
Alarm intelligence
- Alarm classification
pages/1_Realtime.py - Trained classifier model
models/windsense_rf_model.pkl.gz - Model training
utils/model_trainer.py - Anomaly detection
utils/anomaly_detector.py - Alarm prioritization
pages/1_Realtime.py - Root-cause analysis
pages/1_Realtime.py - Fleet and loss analytics
pages/1_Realtime.py
Operations workflow
- Acknowledgment workflow
pages/1_Realtime.py - Email notifications
utils/email_queue.py - SMS notifications
utils/sms_sender.py
Data and state
- SCADA and alarm data
data - Anomaly review log
data/anomaly_log.json - Operational records
utils/database.py
Other components
- Wind-farm operator
- Email service
- SMS provider
Connections 20
- Wind-farm operator → Streamlit entry: opens
- Streamlit entry → Login and signup: routes unauthenticated
- Streamlit entry → Realtime dashboard: routes authenticated
- Login and signup → Realtime dashboard: redirects authenticated
- SCADA and alarm data → Realtime dashboard: supplies historical data
- OPC UA simulator → Realtime dashboard: simulates live readings
- Realtime dashboard → Alarm classification: classifies alarms
- Alarm classification → Trained classifier model: loads model
- Model training → Trained classifier model: trains and saves
- Realtime dashboard → Anomaly detection: analyzes alarms
- Anomaly detection → Anomaly review log: writes review entries
- Realtime dashboard → Alarm prioritization: prioritizes alarms
- Realtime dashboard → Root-cause analysis: supports investigation
- Realtime dashboard → Fleet and loss analytics: presents analytics
- Realtime dashboard → Acknowledgment workflow: handles acknowledgments
- Acknowledgment workflow → Operational records: records alarm actions
- Realtime dashboard → Email notifications: dispatches email notices
- Realtime dashboard → SMS notifications: dispatches SMS notices
- Email notifications → Email service: sends email
- SMS notifications → SMS provider: sends SMS