Chrgnet

Chrgnet is a micro-service application that analyze and predicts behaviours of electric vehicle charinging stations. Chrgnet scrapes live data from charging stations and produce prediction models based on time-series analysis and artifical intelligence to analyze the pressure of stations based on spatio-temporal features. Chrgnet offers features such as:

  • Pressure prediction via time-series analysis and autoencoders
  • Price optimization via integer linear programming optimization

Chrgnet utilize a hexagonal hierarchical spatial index to partitioning the Earth's surface into hexagonal cells inorder to efficiently search in geo-spatial data. The figure below shows charging stations in the south of Sweden togheter their hexagonal grid cells.

Using time-series analysis and autoencoders, the pressure level on charging stations is predicted. The pressure forecasting models are used for prize optimization using integer linear programming optimization. The figure below demonstrates the pressure level at charging stations in the south of Sweden. The color of each cell indicates the utilization of charging stations in the area.

Chrgnet is build using a CI/CD pipeline including Jenkins, Helm and Kubernetes.