Real-time Industrial Equipment Monitoring with AI-Powered Predictive Maintenance
Prevent equipment failures, save lives, reduce downtime. Get alerts in 3 seconds, not hours.
- Industrial equipment fails unexpectedly causing production stops and millions in losses
- Offshore platforms experience fires and explosions - people die because alerts come too late
- Existing solutions (AVEVA PI Vision, IBM Maximo) cost $50,000-500,000/year and take 6-12 months to deploy
PredictiveMaint AI monitors your equipment in real-time and predicts failures before they happen:
- Instant Alerts: Gas leak detected? Alert in 3 seconds, not hours
- Predictive Maintenance: "Compressor bearing will fail in 47 hours - schedule maintenance"
- 10x Cheaper: Enterprise pricing starts at $500/month, not $50,000/year
- Deploy in Days: Not months
This open-source demo includes:
| Feature | Description |
|---|---|
| Live Dashboard | Real-time sensor monitoring with WebSocket updates |
| 10 Sensor Types | Temperature, Vibration, Pressure, Gas, Flow, and more |
| RUL Prediction | Remaining Useful Life countdown with confidence scores |
| Instant Alerts | Critical alerts triggered within seconds |
| Dark Industrial Theme | Professional UI designed for control rooms |
| Simulated Data | Realistic sensor patterns for demonstration |
cd backend
python -m venv venv
venv\Scripts\activate # Linux/Mac: source venv/bin/activate
pip install -r requirements.txt
copy .env.example .env # Linux/Mac: cp .env.example .env
python -m uvicorn app.main:socket_app --reload --port 8000cd frontend
npm install
npm run devOpen http://localhost:5173 in your browser.
| Sensor | Unit | Normal | Warning | Critical | Industries |
|---|---|---|---|---|---|
| Temperature | C | <70 | 70-85 | >85 | All |
| Vibration | mm/s | <2.0 | 2.0-4.0 | >4.0 | All |
| Pressure | PSI | 45-55 | 55-60 | >60 | Oil/Gas |
| Gas Concentration | ppm | 0-10 | 10-100 | >100 | Oil/Gas, Mining |
| Rotation Speed | RPM | 1500-1700 | 1700-1800 | >1800 | Manufacturing |
| Power Consumption | kW | <14 | 14-17 | >17 | All |
| Acoustic Level | dB | <75 | 75-85 | >85 | Manufacturing |
| Oil Quality | % | >70 | 50-70 | <50 | Aviation |
| Flow Rate | L/min | 8-12 | 5-8 | <5 | Oil/Gas |
| Humidity | %RH | 30-60 | 60-75 | >75 | Data Centers |
Frontend:
- React 18 + TypeScript + Vite
- TailwindCSS (dark industrial theme)
- Socket.IO (real-time updates)
- Recharts (data visualization)
- Zustand (state management)
Backend:
- Python 3.12 + FastAPI
- Socket.IO (WebSocket)
- Pydantic (validation)
- SQLite (demo database)
| Method | Endpoint | Description |
|---|---|---|
| GET | /api/health |
Health check |
| GET | /api/machines |
All machine statuses |
| GET | /api/machines/{id} |
Single machine details |
| GET | /api/alerts |
Active alerts |
| GET | /api/alerts/critical |
Critical alerts only |
| GET | /api/predictions |
RUL predictions |
| GET | /api/dashboard |
Full dashboard data |
| POST | /api/simulate/anomaly/{id} |
Inject test anomaly |
| Industry | Use Case |
|---|---|
| Oil & Gas | Monitor compressors, turbines, pipelines, gas detectors |
| Manufacturing | Track CNC machines, robots, conveyors, presses |
| Power Generation | Monitor turbines, generators, transformers |
| Mining | Track excavators, drills, ventilation systems |
| Data Centers | Monitor cooling, UPS, server health |
| Aviation | Engine health, hydraulics, landing gear |
| Water/Wastewater | Pumps, filters, chemical dosing |
| Food & Beverage | Refrigeration, mixers, packaging lines |
The enterprise version includes features not available in this demo:
| Feature | Demo | Enterprise |
|---|---|---|
| Real sensor integration (OPC-UA, MQTT, Modbus) | - | Yes |
| Multi-tenant (separate companies) | - | Yes |
| User authentication & roles | - | Yes |
| Admin panel | - | Yes |
| Email/SMS alerts | - | Yes |
| Historical data storage (years) | - | Yes |
| Custom thresholds per sensor | - | Yes |
| Equipment hierarchy (Site > Zone > Machine) | - | Yes |
| Time range selection (1h, 1d, 7d, custom) | - | Yes |
| Limit lines on charts (HH, H, L, LL) | - | Yes |
| Data export (CSV, PDF) | - | Yes |
| Scheduled reports | - | Yes |
| SSO integration | - | Yes |
| On-premise deployment | - | Yes |
| SLA & support | - | Yes |
| Tier | Sensors | Users | Price |
|---|---|---|---|
| Starter | Up to 100 | 5 | $500/month |
| Professional | Up to 500 | 20 | $2,000/month |
| Enterprise | Unlimited | Unlimited | Custom |
Compare to competitors: AVEVA PI Vision costs $50,000-500,000/year. We're 10x cheaper.
Interested in the enterprise version?
- Email: [Coming soon]
- GitHub Issues: Open an issue for questions
predictive-maintenance-ai/
├── frontend/ # React dashboard
│ ├── src/
│ │ ├── components/ # UI components
│ │ │ ├── dashboard/ # Dashboard, MachineCard, StatsCard
│ │ │ ├── sensors/ # SensorCard
│ │ │ ├── predictions/ # RULClock
│ │ │ └── alerts/ # AlertPanel
│ │ ├── hooks/ # useWebSocket
│ │ ├── stores/ # Zustand state
│ │ ├── lib/ # Socket, utils, API
│ │ └── types/ # TypeScript types
│ └── package.json
├── backend/ # FastAPI server
│ ├── app/
│ │ ├── api/ # REST & WebSocket routes
│ │ ├── models/ # Pydantic schemas
│ │ └── services/ # Business logic
│ │ ├── sensor_simulator.py
│ │ ├── alert_service.py
│ │ └── prediction_service.py
│ └── requirements.txt
└── docs/ # Documentation
Real disasters that better monitoring could have prevented:
- Piper Alpha (1988): 167 workers died - gas leak not detected in time
- Deepwater Horizon (2010): 11 workers died - pressure anomaly missed
- Texas City Refinery (2005): 15 workers died - level indicator failed
Every second counts. Our system alerts in 3 seconds, not 3 hours.
- Real-time dashboard with WebSocket
- 10 sensor types with thresholds
- RUL prediction with ML
- Instant alert system
- Dark industrial theme
- User authentication
- Multi-tenant architecture
- Real sensor integration (OPC-UA, MQTT)
- Historical data & trends
- Mobile app
Contributions welcome! Please read our contributing guidelines first.
MIT License - see LICENSE file for details.
Built with the mission to prevent industrial disasters and save lives.