A 36-month research project building an AI-powered platform that correlates technical, meteorological, socio-economic and environmental data to transform Croatia’s electricity distribution system from reactive to predictive.

1 Oct 2025

Project
start

30 Sep 2028

Project
end

€2,979,261

Total project
value

€2,567,943

Grant
amount

Europe’s energy transition is at a pivotal moment. The rapid electrification of transport, heating and industry – combined with the mass integration of renewable energy sources, storage and prosumers – has created unprecedented strain on electricity distribution systems. Distribution System Operators must manage increasingly complex, decentralised, data-intensive networks while maintaining reliability and affordability.

Croatia’s national DSO, HEP ODS, faces the same systemic challenges: a vast, geographically diverse and technologically heterogeneous network. Although equipped with SCADA, GIS and smart-metering systems, these tools operate largely in silos, limiting the ability to generate predictive insight – leading to delays in renewable connections, higher energy losses, longer outages and suboptimal investment planning.

AIOptiGrid’s mission is to enable the digital and green transformation of electrical distribution systems by unlocking the untapped value of data through advanced analytics, machine learning and artificial intelligence.

Its core objective is to develop and validate a novel software solution that transforms reactive, silo-based management into a proactive, predictive and integrated system – empowering DSOs to make faster, data-driven and more sustainable decisions.

Unlike existing commercial smart-grid tools, the project introduces a cross-disciplinary, context-aware methodology that integrates technical and operational grid data, meteorological and climatological variables, socio-economic and demographic indicators, and environmental and market data – revealing dependencies invisible to conventional planning, such as how weather, urbanisation or consumer behaviour shape grid performance.

The full implementation plan runs across the project’s entire duration, structured so research results transfer into a viable product at TRL 7–8.

Infrastructure Equipping M1–M6

Installation of high-performance GPU servers, data lake and secure networking infrastructure.

Research Activities M1–M36

Results-oriented research structured for transfer into a viable product, across four sub-activities:

  • Data Integration & Modeling (M1–M18) — collection, anonymisation, cleaning and standardisation of multi-source data.
  • AI & Advanced Analytics Development (M6–M24) — deep learning, time-series forecasting, regression and transformer-based models to identify correlations and predict network behaviour.
  • Prototype Development & Validation (M12–M30) — software prototype, visualisation tools and decision-support interfaces, validated on real-world datasets.
  • Business Case Definition (M18–M36) — at least three validated use cases (loss reduction, outage prediction, investment optimisation) with quantifiable performance indicators.

Knowledge & Technology Transfer M1–M36

Integration of research outputs into ADNET’s commercial NETVISION platform, with free operational deployment at HEP ODS.

Capacity Building & Dissemination M12–M36

Training, workshops and international presentations to spread knowledge and prepare for broader EU market replication.

Market Analysis & Business Validation M1–M12

A comprehensive study covering market potential, competition, SWOT and risk analysis, and commercial viability.

Promotion & Visibility M1–M36

Communication with professionals and the public via social media and a dedicated project website.

Project Management M1–M36

Day-to-day coordination, partner support, progress tracking and financial reporting.

AIOptiGrid is built for the operational realities of medium-sized DSOs, and goes further.

01

NLP-derived AI models

Applies AI methods originally developed for natural language processing to detect hidden dependencies in energy datasets.

02

Non-technical data fusion

Integrates climate, demographic and urban-development data for context-aware decision-making – dimensions most systems overlook.

03

Explainable by design

A transparent, explainable AI framework that lets DSOs understand and trust model outputs, not just consume them.

AIOptiGrid is built for the operational realities of medium-sized DSOs, and goes further.

Greater reliability, shorter outages, faster renewable connections and lower long-term electricity costs.

Improved asset utilisation, operational efficiency and flexibility-service integration.

A replicable, evidence-based model for the digital transformation of energy networks across the EU.

New methodologies for multidisciplinary data analytics and AI in critical infrastructure.