The Need
A joint venture of a European research organization and a Swedish company requested Gadgeon to industrialize the peak saving and balancing mechanism for smart city heat grids.
The platform to collets temperatures from buildings & networks, and calculate the next day energy need per building, considering production capacity of the grid, thermal mass and occupancy of the building.
Impacts Delivered
- Considerable reduction in the peak load of the heating grid that in turn reduced the cost of operation considerably.
- Significant reduction in usage of fossil fuels and carbon emission
- Easy Personalized Dynamic Visualization for monitoring and analyzing time series data.
Our Solution
- Gadgeon architected the platform for hosting Python Algorithms in Azure cloud.
- Algorithms for peak shaving using a forecaster, planner and tracker method.
- Solution involved setup of ARM template which can deploy entire solution for clients customers
- Solution will have Configuration portal to configure required parameters for processing peak shaving algorithm.
- Integration of PowerBI for creating dashboards for analytics and performance
- Alerting mechanisms for non-performance
Cloud Application
- Microservice architecture achieved using Dockers and Kubernetes
- Azure SQL server used as semantic layer for power Bi reporting
- NOSQL CosmosDB for metadata storage. Azure data lake used for Raw ad curated data
- ARM template for packaging & deployment
- Training the model using historical data, and then forecasting the peak demand and the variation in temperature set points required to achieve peak saving.
Tools & Technologies
- Cloud server hosted in Azure,
- React JS based web application
- Embedded Power BI for reporting using Azure SQL server
- Azure IoT Hub for communication with Gateway influencer
System Architecture Diagram
Sample Prediction Outcomes
