The transformation of the electricity system is not only about integrating more renewable energy. It is also about managing a much more dynamic and complex distribution grid, where generation and consumption patterns are increasingly variable and decentralized. In this context, data becomes a key asset for enabling smarter and more resilient grid operation.
Within the European project PISTIS, the Energy Demonstrator explores how secure and controlled data sharing can support new services for the operation of distribution networks. The demonstrator focuses on improving grid resilience by leveraging the flexibility offered by prosumers, coordinated through aggregators and market platforms. This creates a real environment where data flows between different actors — DSOs, aggregators, market operators and technology providers — can be tested and validated.
A key aspect of this demonstrator is not only the availability of data, but also how it is processed and transformed into actionable insights. Through the PISTIS platform, datasets from different sources are ingested, enriched, checked, and made interoperable. These datasets include grid topology, historical consumption and generation, distributed energy resources (DERs), and external information such as meteorological data. Once processed, they can be used by analytics services and shared again with other stakeholders in a controlled and traceable way.
In this data-driven ecosystem, CARTIF contributes with an event detection and prediction algorithm, developed as part of the hosting capacity use case. The role of this algorithm is to anticipate potential problems in the distribution grid — such as congestion or voltage issues — before they actually occur.
The algorithm operates by combining two main types of input data. On the one hand, it uses static information about the network, such as its topology and technical constraints. On the other hand, it relies on forecasted demand at each node of the grid, which can be enhanced with external data sources like weather information (provided by UBIMET). Based on these inputs, the algorithm solves a non-linear optimisation problem to evaluate whether the predicted operating point is compatible with safe grid operation.
Instead of simply detecting problems after they happen, the algorithm answers two practical questions: will a grid event occur? and what changes would be needed to avoid it? To do so, it searches for the minimum adjustment required in demand across the network to keep the system within its operational limits. If the required adjustment exceeds a predefined threshold at any node, the algorithm flags the situation as a potential event.
This approach provides not only a binary indication (event / no event), but also a quantitative estimation of the flexibility needed and its location in the network. This information is especially valuable within the PISTIS ecosystem, as it feeds other components of the demonstrator.
More specifically, the outputs generated by CARTIF are shared through the PISTIS platform with CUERVA, which enriches them with additional information and shares them with BAMBOO. These results can then be used to trigger flexibility actions, for example by launching a flexibility request in the market or activating distributed energy resources in specific grid zones.
In this sense, the algorithm plays a clear role within the overall workflow of the Energy Demonstrator. It sits in the analytics layer, consuming structured and enriched data from the platform, and producing new data assets that support decision-making processes. These outputs are fully traceable within PISTIS, ensuring transparency in how data is used and exchanged between stakeholders.
The results obtained during the validation phase have shown that combining data sharing mechanisms with advanced analytics can significantly improve the anticipation and management of grid events. More importantly, they demonstrate how different actors — from DSOs to aggregators and market operators — can collaborate through data to enable more flexible and efficient grid operation.
By integrating predictive algorithms like this into a broader data ecosystem, PISTIS supports the development of more digital, interoperable and resilient energy systems.
