Advanced Predictive Maintenance Platform

A modern, intelligent web application for predictive maintenance in manufacturing environments. This system uses machine learning to predict equipment failures and provides real-time monitoring through an interactive dashboard with specific failure detection and immediate service request capabilities.

🚀 Key Features

• 🎯 Specific Failure Detection: Shows exact issues like "Nozzle Clog", "Pump Wear", "Heater Element Degradation" instead of generic messages
• ⚡ Immediate Service Requests: Click "Request Service" → Select specific parts → Get instant confirmation with time estimates
• 📊 Interactive Dashboard: Real-time machine health monitoring with clickable sensor graphs
• 🤖 Intelligent ML System: Supports different machine types with varying sensor configurations
• 📈 Predictive Analytics: Failure prediction and remaining useful life estimation
• 🔄 Real-time Updates: Live data streaming with WebSocket support
• 📱 Responsive Design: Mobile-friendly sidebar with dropdown navigation
• 🚨 Anomaly Detection: Statistical analysis with configurable thresholds
• 🔧 Maintenance Scheduling: Professional service request workflow

🛠 Technology Stack

• Backend: Flask 3.0 with SQLAlchemy
• Machine Learning: TensorFlow 2.15 with dual-output GRU models
• Frontend: Modern JavaScript with Chart.js and Tailwind CSS
• Real-time: Flask-SocketIO for live updates
• Database: SQLite (PostgreSQL ready)
• Architecture: Centralized package management with clean structure

⚡ Quick Start

1. Install Dependencies

pip install -r requirements.txt

2. Initialize Database

python -c "from app import create_app; app = create_app(); print('✅ Database initialized')"

3. Train Models (Optional - system works with intelligent fallback)

python scripts/create_basic_models.py

4. Run Application

python run.py

5. Access Dashboard

Open http://localhost:5000 in your browser

🏗 Project Structure

PDM/
├── app/
│   ├── packages.py          # Centralized package definitions
│   ├── __init__.py         # Application factory
│   ├── api/                # API endpoints
│   │   ├── routes.py       # Main API routes
│   │   └── machine_management.py
│   ├── config/             # Configuration files
│   │   └── machine_config.py
│   ├── models/             # Database models
│   │   ├── company.py
│   │   └── sensor_data.py
│   ├── services/           # Business logic
│   │   ├── predictor.py    # ML prediction service
│   │   └── simulator.py    # Data simulation
│   └── templates/
│       └── dashboard.html  # Interactive dashboard
├── scripts/                # Utility scripts
├── models/                 # Trained ML models
├── .gitignore             # Git ignore rules
├── requirements.txt       # Python dependencies
└── run.py                # Application entry point

🎯 Specific Failure Detection

The system provides exact failure information instead of generic messages:

Sensor-Based Failures (Immediate Detection)

• FILLING: "Nozzle Clog", "Pump Pressure Fault", "Bearing Wear"
• SHRINK: "Heater Failure", "Conveyor Jam", "Humidity Sensor Fault"
• PILLOW: "Seal Heater Failure", "Pressure Drop", "Alignment Error"
• BOXING: "Pneumatic Leak", "Box Jam", "Feed Mechanism Fault"

ML-Based Predictions (For Degrading Machines)

• FILLING: "Nozzle Clog", "Pump Wear", "Seal Deterioration"
• SHRINK: "Heater Element Degradation", "Temperature Sensor Drift", "Insulation Failure"
• PILLOW: "Seal Heater Wear", "Pressure Regulator Drift", "Jaw Misalignment"
• BOXING: "Pneumatic Seal Wear", "Conveyor Belt Stretch", "Sensor Calibration Drift"

⚡ Immediate Service Request System

Professional Workflow:

1. Click "Request Service" on any machine
2. Select specific parts from machine-specific list
3. See time estimates for each part (15-70 minutes)
4. Get instant confirmation with notification
5. Track service progress with real-time updates

Serviceable Parts by Machine Type:

FILLING Machines

• Nozzle Assembly (30 min) - Clean/replace nozzles
• Pump System (45 min) - Service pump and seals
• Fill Volume Sensor (15 min) - Calibrate/clean sensor
• Vibration Sensor (10 min) - Check mounting and calibration
• General Maintenance (60 min) - Full system inspection

SHRINK Machines

• Heating Elements (40 min) - Replace/repair heating elements
• Conveyor Belt (35 min) - Adjust/replace belt
• Temperature Sensor (20 min) - Calibrate temperature sensor
• Humidity Control (25 min) - Service humidity system
• General Maintenance (70 min) - Full system inspection

PILLOW Machines

• Sealing Jaws (50 min) - Replace/align sealing jaws
• Pressure System (30 min) - Service pneumatic system
• Alignment Mechanism (40 min) - Calibrate alignment system
• Temperature Control (25 min) - Service heating system
• General Maintenance (65 min) - Full system inspection

BOXING Machines

• Pneumatic System (35 min) - Service air compressor and lines
• Conveyor Belt (30 min) - Adjust/replace conveyor
• Box Feed Mechanism (45 min) - Service box handling system
• Pressure Sensors (15 min) - Calibrate pressure sensors
• General Maintenance (60 min) - Full system inspection

📊 Interactive Dashboard Features

Real-time Monitoring

• Machine Cards: Live sensor readings with color-coded status
• Clickable Charts: Interactive sensor graphs with threshold indicators
• Status Badges: Visual indicators for Normal/Degrading/Maintenance states
• RUL Display: Remaining Useful Life predictions in hours

Responsive Sidebar

• Desktop: Traditional tabs (Watchlist, Event Log)
• Mobile: Dropdown menu with icons for better usability
• Watchlist: Critical failures with specific part information
• Event Log: Real-time system events and maintenance history

Professional Notifications

• Success Notifications: Green notifications for completed actions
• Error Notifications: Red notifications for failed operations
• Service Confirmations: Detailed service request confirmations

🔧 API Endpoints

Endpoint	Method	Description
/	GET	Main dashboard
/dashboard	GET	Interactive dashboard
/api/latest-states	GET	Current machine states and predictions
/api/history	GET	Historical sensor data with charts
/api/request-maintenance	POST	Professional service request
/api/machine-mgmt/*	GET/POST	Machine management operations

⚙️ Configuration

Environment Variables (Optional - sensible defaults provided)

SECRET_KEY=your-secret-key
DATABASE_URL=sqlite:///maintenance.db
MODEL_BASE_DIR=./models
SIMULATOR_UPDATE_INTERVAL=1
WINDOW_SIZE=24
ANOMALY_Z_SCORE_THRESHOLD=4.0

Machine Configuration

Add new machine types in app/config/machine_config.py:

MACHINES = {
    'YOUR_MACHINE': {
        'sensors': ['Sensor1', 'Sensor2', 'Sensor3'],
        'failures': ['Failure1', 'Failure2', 'Failure3']
    }
}

🧹 Project Cleanup

Removing pycache Directories

The project includes comprehensive cleanup tools:

# Remove all __pycache__ directories
Get-ChildItem -Path . -Recurse -Directory -Name "__pycache__" | ForEach-Object { Remove-Item -Path $_ -Recurse -Force }

# Or use Python
python -c "import os; [os.system(f'rm -rf {root}/{dir}') for root, dirs, files in os.walk('.') for dir in dirs if dir == '__pycache__']"

.gitignore Configuration

The project includes a comprehensive .gitignore file that prevents:

• __pycache__/ directories
• Python compiled files (*.pyc, *.pyo)
• Environment files (.env, venv/)
• IDE files (.vscode/, .idea/)
• OS files (.DS_Store, Thumbs.db)
• Database files (*.db, *.sqlite)
• ML model files (*.h5, *.pkl)

🚀 Production Deployment

Docker Deployment

docker build -t maintenance-platform .
docker run -d -p 5000:5000 maintenance-platform

Environment Setup

export FLASK_ENV=production
export DATABASE_URL=postgresql://user:pass@localhost/db
python run.py

Performance Optimization

• Debug Mode: Disabled for production (debug=False)
• Connection Pooling: Configured for high-traffic environments
• Model Caching: ML models loaded once and cached
• Database Indexing: Optimized queries for sensor data

🔄 Continuous Learning

The system supports continuous learning from new data:

• Automatic Model Updates: Models retrain on new sensor data
• Failure Pattern Learning: System learns from new failure modes
• Performance Adaptation: Thresholds adjust based on historical data

🛡 Security Features

• Input Validation: All API inputs validated and sanitized
• SQL Injection Protection: SQLAlchemy ORM prevents SQL injection
• CORS Configuration: Properly configured for production
• Error Handling: Comprehensive error handling without information leakage

📈 Performance Metrics

• Response Time: < 100ms for dashboard updates
• Memory Usage: Optimized for 24/7 operation
• Database Performance: Indexed queries for fast sensor data retrieval
• ML Inference: < 50ms per prediction

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

Development Guidelines

• Follow PEP 8 style guidelines
• Add comprehensive docstrings
• Include unit tests for new features
• Update documentation for API changes

📄 License

MIT License - see LICENSE file for details.

🆘 Troubleshooting

Common Issues

Server won't start:

# Check if port 5000 is available
netstat -an | findstr :5000

# Kill existing Python processes
taskkill /f /im python.exe

Database errors:

# Recreate database
rm maintenance.db
python -c "from app import create_app; app = create_app()"

ML model errors:

# Recreate basic models
python scripts/create_basic_models.py

pycache cleanup:

# Remove all cache directories
Get-ChildItem -Path . -Recurse -Directory -Name "__pycache__" | Remove-Item -Recurse -Force

📞 Support

For issues and questions:

1. Check the troubleshooting section above
2. Review the API documentation
3. Check the console for error messages
4. Ensure all dependencies are installed correctly

---

Built with ❤️ for modern manufacturing environments