Non-disruptive energy data thanks to AI that improves renewable energy monitoring

Ensuring continuity and reliability in data is essential for those operating in the renewable energy market. This practical need led to the collaboration between Fondazione Bruno Kessler and the South Tyrolean company Tecno.Energy, part of the broader Psaier.Energies group, an energy-sector organization offering integrated services across the renewable energy, utilities, and smart metering supply chain.  Tecno.Energy uses data loggers, sensors, and probes installed on systems to monitor electricity production and delivery to the grid in near real time.

For this reason, one of their main operational challenges was interruptions lasting several minutes or hours in data transmission, which risked compromising monitoring quality and operational decision-making, and led to increasingly frequent manual interventions to resolve anomalies.

The collaboration with the Data and Knowledge Management (DKM) unit at the FBK Center for Augmented Intelligence, led by Luciano Serafini, focused specifically on this issue. The work was divided into two complementary tracks.  On one hand, the nature of the interruptions was analyzed to identify their causes. On the other, artificial intelligence models were developed to estimate missing data when recovery was not possible in time.

In the first track, the FBK team analyzed system operations by developing a model capable of automatically classifying errors, distinguishing between hardware and software issues and categorizing them. This made it faster to identify causes, allowing the company to quickly determine, for example, whether a probe was disconnected or if there was a system anomaly. In particular, software errors fell into recurring categories, making it possible to introduce targeted controls and suggest improvements. The result was a more efficient error management process, with a significant reduction in the time needed to identify issues and a corresponding decrease in their direct impact on daily operations.

In the second track, the focus was on cases where data could not be recovered, implementing specific predictive models: some developed and trained by the FBK team for the specific context, others based on established time-series analysis techniques. The contribution was to bring in and make accessible already mature artificial intelligence solutions, adapting them to the operational context and integrating them into existing systems.
Comparing these models, adapted and validated against Tecno.Energy’s operating conditions and historical data, made it possible to identify the best-performing solution for reliably estimating missing values.

The project, carried out in the first quarter of 2026, is part of the broader DIPS – Digitalization and Innovation of Public Services initiative. Led by FBK and awarded the European Commission’s “Seal of Excellence,” this initiative aims to accelerate the digital transformation of SMEs and public administration through the adoption of advanced technologies such as artificial intelligence.

The collaboration between FBK and Tecno.Energy demonstrates how artificial intelligence can be applied concretely to energy monitoring—not as an abstract concept, but as an operational tool to improve data reliability, support anomaly diagnosis, and manage information gaps in measurement systems.