At ENLIT Europe 2025, OPENTUNITY once again demonstrated its commitment to shaping the future of smart grid innovation. This year, the spotlight was on the work of Elektro Ljubljana, whose contribution showcased how artificial intelligence can dramatically improve reliability and performance in low-voltage distribution networks, an area where fault detection remains notoriously challenging.

The presentation, delivered by Dr. Klemen Nagode during the session “AI in Practice – Part 2”, offered an insightful look at a rapidly evolving field where data, algorithms, and operational expertise converge to strengthen tomorrow’s electricity distribution systems.

Bringing AI Into the Heart of Low-Voltage Operations

Unlike transmission grids, low-voltage networks are vast, complex, and traditionally less monitored, making it difficult for operators to quickly identify emerging issues. Faults such as bad neutral contacts, voltage anomalies, or early signs of component degradation often remain invisible until they escalate into service interruptions.

To tackle this, Elektro Ljubljana developed an AI-based anomaly detection system designed to uncover these hidden issues early on. The solution analyses a rich blend of data sources, smart meters, protection relays, historical fault records, and state estimators, to detect patterns that signal abnormal behaviour.

Behind the scenes, the team carried out extensive feature engineering, model training, and data validation, using advanced machine learning techniques such as weighted K-nearest neighbors with optimised parameters. The result is a system capable of identifying subtle anomalies with remarkable speed, supporting predictive maintenance and reducing operational uncertainty.

How OPENTUNITY Strengthens This Work

The development does not stand alone, it is actively supported by several tools created within the OPENTUNITY project. Among them:

• Fuse-burn detection modules offering early indicators of device or line distress
• Real-Time Thermal Rating (RTTR), enabling dynamic monitoring of conductor temperature and loading
• Failure forecasting tools, combining short-term and long-term risk modelling
• Advanced asset management systems, providing a structured view of network conditions

These components bring additional layers of insight to the anomaly detection framework, helping enrich datasets, refine machine learning models, and improve the robustness of the overall solution. This illustrates OPENTUNITY’s mission: to build interoperable, complementary tools that elevate grid management through digital intelligence.

A Human–Machine Partnership

One of the strongest points highlighted in the session was the importance of collaboration with field maintenance teams. Real-world verification of model predictions, by technicians working directly on the network, provided essential feedback, allowing the algorithms to evolve beyond theoretical accuracy and into practical, operational usefulness.

This iterative loop between AI systems and human expertise created a powerful validation mechanism, ensuring that the tool performs reliably in diverse, real-life scenarios.

Why This Matters for the Future of Grid Operations

This work marks an important step forward for European distribution system operators. By combining AI, operational data, and domain expertise, the approach significantly improves fault detection capabilities in low-voltage networks, historically one of the most difficult areas to monitor.

It also demonstrates the potential of the solutions being developed in OPENTUNITY: tools that are not only innovative but also applicable, scalable, and designed to support day-to-day decision making.

Download the Presentation

Interested in the technical details?

👉 Download the full ENLIT 2025 presentation slides here:

Robust AI-Based Anomaly Detection for Enhanced Performance in Low-Voltage Distribution Networks. – Opentunity