AI-powered grid assistant goes open source

AINETUS, which was developed within the AI4REALNET project, provides AI components designed to augment decision making in power grid operations through intuitive interfaces and visualisations that allow operators to assess system conditions and AI recommendations immediately. 

As such, it is considered to offer a new paradigm for AI-assisted operation of a safety-critical infrastructure. 

With its integration into LF Energy as an open source solution, it should contribute to the development of transparent and scalable AI technologies for grid operation use cases and facilitate the industry adoption of AI-based assistants in their operations. 

It also should extend the impact of the tools by ensuring they remain accessible, actively maintained, and capable of generating long-term impact beyond the official conclusion of the AI4REALNET project in March 2027. 

Also of interest
UK Power Networks’ DSO launches open data analysis package
Portugal grid constraints hamper renewable energy deployment

Ricardo Bessa, a researcher at Portugal’s INESC TEC and head of the Power & Energy Systems domain, which led the development, said that this cooperation represents an important step in advancing AI-based assistants for power system operation. 

“It also reinforces the project’s commitment to open science and open innovation, ensuring that key software assets remain accessible, actively maintained and capable of generating long-term impact beyond the duration of EU-funded research projects.” 

AI4REALNET is focussed on the development of AI-based solutions addressing three critical systems – electricity, railways and air traffic management – modelled by networks that can be simulated and are traditionally operated by humans, but where AI systems complement and augment the human abilities. 

Its core elements are AI algorithms composed mainly by supervised and reinforcement learning, human-in-the-loop decision making for co-learning between AI and humans, and autonomous AI systems relying on human supervision. 

The basis for AINETUS was two electricity network use cases, one overseeing the transmission grid using SCADA data and energy management system tools to identify and manage congestion, and the second to transfer the AI assistant from simulation to real world operation through two pathways to managing the transmission grid, i.e. in coping with real world conditions and alerting the operator when data limitations prevent full autonomy. 

Within the collaboration with LF Energy, selected project outcomes will be embedded into established open-source ecosystems and governed under the LF Energy framework, ensuring transparent development, community governance and long-term maintenance.  

The open source initiative will integrate key technological building blocks developed within AI4REALNET, including reinforcement learning-based agents for operational decision support, high performance power flow solvers for real time analysis, uncertainty quantification modules for risk aware operation and human-AI interaction capabilities based on the hypervision concept. 

By hosting AINETUS within LF Energy, the platform is decoupled from single organisation ownership and embedded within an ecosystem that supports open development, collaborative innovation and technology transfer across academia, industry and infrastructure operators. 

In addition to the Horizon Europe support, AI4REALNET received funding from the government of Switzerland.