Skip to content

Repository files navigation

NetZero

CI

Overview

NetZero is a hybrid Edge–Cloud Software-as-a-Service (SaaS) platform that bridges the gap between thermodynamic building modelling and real-time electricity grid performance. By combining machine learning-based thermal load prediction with live carbon intensity forecasts, the platform enables carbon-aware operational scheduling that reduces energy costs, lowers carbon emissions, and improves sustainability outcomes.


Problem Statement

Buildings often consume energy without considering the carbon intensity of the electricity supplying the grid. This leads to:

  • Inefficient operation of HVAC and other flexible assets.
  • The Demand Trap, where energy-intensive devices operate during periods of high carbon intensity.
  • Increased operational costs.
  • Difficulty meeting sustainability and ESG targets.

NetZero addresses these challenges by integrating physical building characteristics with real-time and forecasted grid carbon intensity, enabling intelligent load shifting, automated decision-making, and carbon-aware operational planning.


Target Users

Homeowners

Reduce inefficient HVAC cycling by leveraging thermodynamic predictions to maintain comfortable indoor temperatures while minimizing energy consumption through informed ventilation and shading strategies.

Small Business Owners

Reduce operational carbon emissions by scheduling flexible energy-intensive activities, such as pre-heating equipment or refrigeration cycles, during periods of lower grid carbon intensity.

Facility Managers

Achieve organizational sustainability and ESG objectives through automated threshold-based load modulation that dynamically adjusts non-critical assets during carbon-intensive periods.


Key Features

  • Thermodynamic Digital Twin Onboarding
  • Building Profile Management
  • Flexible Asset Registration
  • Regional Carbon Forecast Ingestion
  • Carbon-Aware Operational Scheduling
  • Predictive Grid Intensity Analytics
  • Automated Threshold-Based Load Modulation
  • Portfolio Dashboard
  • ESG Reporting
  • Retrofit Simulation
  • Sustainability Report Generation
  • Email Notifications
  • Authentication and User Management

Technology Stack

Frontend

  • Next.js
  • TypeScript
  • Tailwind CSS
  • shadcn/ui
  • Recharts

Backend

  • Python
  • Django REST Framework

Machine Learning

  • PyTorch
  • CoreML

Infrastructure

  • PostgreSQL (NeonDB)
  • Render
  • Vercel
  • GitHub Actions (CI/CD)

Data Sources

UCI Energy Efficiency Dataset

Used to train machine learning models that predict heating and cooling loads using eight structural building parameters.

Repository: https://archive.ics.uci.edu/ml/datasets/Energy+efficiency

Structural Parameters

  • Relative Compactness
  • Surface Area
  • Wall Area
  • Roof Area
  • Overall Height
  • Orientation
  • Glazing Area
  • Glazing Area Distribution

Outputs:

  • Heating Load
  • Cooling Load

NESO Carbon Intensity API

Used to retrieve real-time and forecast carbon intensity information for regional electricity grids, enabling carbon-aware scheduling and automated energy management.

https://carbonintensity.org.uk/


System Architecture

                    +----------------------+
                    |     Next.js UI       |
                    |  Dashboard & Forms   |
                    +----------+-----------+
                               |
                               |
                    Django REST Framework
                               |
        +----------------------+----------------------+
        |                      |                      |
        |                      |                      |
 Building Services      ML Inference Engine    Carbon Forecast Engine
        |                      |                      |
        +----------------------+----------------------+
                               |
                         PostgreSQL (NeonDB)
                               |
                  Scheduled Tasks & Automation

Running the Application Locally

1. Clone the Repository

git clone https://github.com/dr1810/NetZero.git
cd NetZero

2. Start the Frontend

Open a new terminal.

cd netzero-frontend
npm install
npm run dev

The frontend will normally be available at:

http://localhost:3000

3. Start the Backend

Open a second terminal.

cd netzero-backend
pip install -r requirements.txt
python manage.py migrate
python manage.py runserver

The backend will normally be available at:

http://127.0.0.1:8000

4. Open the Application

Navigate to

http://localhost:3000

Continuous Integration

Every push and pull request automatically triggers the GitHub Actions pipeline.

The pipeline performs:

  • Backend validation
  • Frontend build verification
  • Unit testing
  • Integration testing
  • Dependency installation
  • Continuous deployment readiness checks

Pipeline status:

https://github.com/dr1810/NetZero/actions


Project Structure

NetZero
│
├── netzero-frontend
│   ├── app
│   ├── components
│   ├── lib
│   ├── public
│   └── styles
│
├── netzero-backend
│   ├── api
│   ├── config
│   ├── services
│   ├── models
│   ├── migrations
│   └── tests
│
├── .github
│   └── workflows
│
├── docs
├── README.md
└── LICENSE

Machine Learning

The platform predicts a building's thermal behaviour using a neural network trained on the UCI Energy Efficiency Dataset.

The prediction engine estimates:

  • Heating Load
  • Cooling Load

These predictions are combined with live regional carbon forecasts to generate optimized operational schedules.


Carbon-Aware Optimization

The optimization engine combines:

  • Building thermodynamic characteristics
  • Predicted heating and cooling demand
  • Regional carbon intensity forecasts
  • User-defined sustainability preferences

The system then:

  • Identifies optimal operating windows
  • Generates carbon-aware schedules
  • Recommends low-carbon operating periods
  • Automatically modulates eligible assets when thresholds are exceeded

Testing

The project includes:

  • Backend API Tests
  • Integration Tests
  • Machine Learning Validation
  • Frontend Validation
  • GitHub Actions CI Pipeline

Live Deployment

Frontend

https://net-zero-black.vercel.app/

Backend API

https://netzero-jonf.onrender.com/


Repository

GitHub Repository

https://github.com/dr1810/NetZero


Release

Latest Release

https://github.com/dr1810/NetZero/releases/tag/v1.0.0


Future Work

Future development of NetZero will focus on:

  • Advanced reinforcement learning for adaptive scheduling
  • Occupancy-aware optimization
  • Smart home and IoT device integration
  • Battery and renewable energy optimization
  • Mobile application support
  • Multi-region carbon forecasting
  • Enhanced ESG analytics
  • Explainable AI for scheduling decisions

Authors

Dharun Ramesh

New York University

M.S. Computer Science

GitHub: https://github.com/dr1810


License

This project was developed for academic purposes as part of the Software Engineering I course at New York University.

About

Proactive Carbon Optimization for Homes and Businesses

Resources

Security policy

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Used by

Contributors

Languages