The power grid of the future created with federated machine learning

Purpose and goal

The goal of the project is to investigate how federated machine learning can be applied to power grid data and what is required for this to be implemented. The project will map opportunities and challenges with this technology for the electricity network companies.

Expected effects and result

The project´s ambition is to clarify how federated machine learning should be applied to power grid data. This means analyzing challenges linked to technology, IT infrastructure, cyber security, etc. to ensure that no data leaks between parties. The goal is that a first ML model can be created to analyze the behavior of different electricity grid customers.

Planned approach and implementation

Two WPs: 1) This work package focuses on selecting and quality checking the data on which the ML model will be trained. It includes an investigation into interfaces connected to the network companies´ internal system tools as well as ensuring various cyber security issues for network data. 2) This work package drives the actual training of the model in order to make the model discover the power profiles of different electricity users and be able to predict the behavior of new electricity users.