ML-Driven Microgrid Optimization Platform

Interactive web application for intelligent microgrid design using machine learning and network optimization algorithms.

🚀 Key Features

• ML-Powered Energy Prediction: Random Forest model predicts building energy demand based on area, type, and weather conditions
• Intelligent Clustering: DBSCAN algorithm groups buildings by geographic proximity (500m radius) and energy consumption patterns
• Network Optimization: Minimum spanning tree algorithms minimize cable costs while ensuring grid connectivity
• Interactive Visualizations: Real-time Bokeh maps display optimized microgrid networks overlaid on downtown Toronto
• Weather Integration: Incorporates extreme weather scenarios for robust grid planning

🚀 Getting Started

Prerequisites

• Python 3.x
• pip package manager

Installation & Setup

# Clone the repository
git clone <repository-url>
cd ml_microgrids

# Install required dependencies
pip install -r requirements.txt

# Run the application
python app.py

Usage

1. Open your web browser and navigate to http://localhost:5000
2. Access the interactive microgrid optimization interface
3. Input building parameters and weather conditions
4. View ML-generated energy predictions and optimized network topology
5. Explore interactive Bokeh visualizations of the microgrid design

📁 Project Structure

ml_microgrids/
├── app.py                          # Main Flask application with ML pipeline
├── requirements.txt                # Python dependencies
├── conda-requirements.txt          # Conda environment specifications
├── Procfile                        # Heroku deployment configuration
├── runtime.txt                     # Python runtime version
├── data/                           # Data files and ML models
│   ├── mlmicrogrid.pkl            # Trained Random Forest model
│   ├── testweather.csv            # Weather data for predictions
│   ├── Downtown Toronto.gml       # Network graph for Toronto case study
│   ├── dowtown_toronto_lat_long.csv # Building coordinates
│   ├── xs.pkl & ys.pkl            # Cached coordinate data
└── templates/                      # HTML templates
    ├── index.html                  # Main input interface
    ├── analysis.html               # Results visualization
    ├── toronto.html                # Toronto case study
    └── about.html                  # Tool information

🛠️ Technical Stack

Backend: Flask, scikit-learn, NetworkX, pandas, NumPy
ML Algorithms: Random Forest, DBSCAN Clustering, PCA
Visualization: Bokeh interactive maps with CartoDB tiles
Geospatial: PyProj coordinate transformations, Haversine distance calculations
Deployment: Heroku-ready with gunicorn

📊 Capabilities Demonstrated

• Machine Learning: Supervised learning for energy demand forecasting with weather feature engineering
• Unsupervised Learning: Density-based spatial clustering for optimal building groupings
• Graph Theory: Network analysis and minimum spanning tree optimization
• Geospatial Analysis: Coordinate system transformations and distance calculations
• Full-Stack Development: End-to-end web application with interactive frontend and ML backend

🎯 Business Impact

Optimizes microgrid infrastructure for urban environments, reducing energy costs and improving grid resilience during extreme weather events through intelligent ML-driven network design.

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

👨‍💻 Author

Akshay Sharma

• Data Scientist
• Expertise in large-scale data processing and machine learning