GridSense AI is an innovative application designed to optimize micro-grid energy management in underserved areas. Leveraging the power of edge computing and advanced AI models, our solution provides real-time insights and recommendations for efficient energy distribution and consumption.
- Data Upload & Analysis: Upload CSV files containing real-world energy consumption data from edge devices.
- Quick Insights: View key statistics including median, minimum, and maximum energy usage for various appliances.
- Interactive Visualizations:
- Explore energy consumption trends through dynamic graphs.
- Watch a real-time animation of energy flow in the micro-grid system.
- AI-Powered Recommendations: Receive intelligent suggestions for energy optimization, load balancing, and cost reduction.
- Edge Computing Simulation: Experience how lightweight AI models can process data locally for immediate decision-making.
- Scalability Projection: Visualize the potential impact of managing multiple micro-grids across a region.
- Frontend/Backend: React/Astro
- AI Models: Llama 3.2 (1B/3B for edge)
- Data Visualization: Recharts
-
Clone the repository: git clone https://github.com/mohdlatif/GridSenseAI.git
-
Install dependencies: cd GridSenseAI npm install
-
Start the development server: npm run astro dev
Important Note: For the best experience in testing the AI capabilities of GridSense AI, we recommend running the application locally using npm astro dev.
The AI analysis typically takes between 90 to 120 seconds to process the data and generate comprehensive recommendations. Due to this extended processing time, the API endpoint may exceed the 60-second limit imposed by Vercel in the deployed version.
By running the app locally, you can experience the full capabilities of the AI analysis without timing out, ensuring you see the complete set of insights and recommendations generated by GridSense AI.
Our dataset is available through two convenient options:
Original Sources: Access the raw data directly from its primary sources. This option is ideal for users who prefer to work with the most up-to-date information or require specific data subsets. Google Drive Repository: For ease of access, we've compiled the complete dataset in a Google Drive folder. This option is particularly useful due to the large file sizes, which exceed GitHub's upload limits.
To obtain the dataset, please choose the method that best suits your needs and technical requirements.
- https://datadryad.org/stash/dataset/doi:10.5061/dryad.fqz612jzb
- https://www.kaggle.com/datasets/yekenot/power-data-from-mesa-del-sol-microgrid
- https://drive.google.com/drive/folders/1SBwyO0N3bsmAtaBNe1IYjvnW-xaLFpyS?usp=sharing
Developed with ❤️ for the Edge Runners 3.2 by powerflow