Ariadna is presenting a data governance-based approach at CIRED Workshop 2026 to help electricity distribution companies turn heterogeneous information into observable, simulable and operable grid models. The proposal combines partial topologies, dispersed systems and measurement data to build more reliable models capable of supporting grid operation, simulation and planning.
The paper, presented under the title “Data Governance for LV Grid Twins. Lessons from Spain and Portugal”, shows how proper data management enables specific use cases for DSOs: feeder and phase identification, localisation of affected customers, support for switching operations, loss analysis, fraud detection, identification of DERs, overvoltage detection and phase balancing, with measurable and auditable results. In deployments in Spain and Portugal, the methodology produced auditable feeder, phase and customer-affectation outputs across 60k secondary substations, 250k LV feeders and +3M smart meters.
Heterogeneous electrical data can be converted into useful operational decisions when it is integrated, validated and contextualised through algorithms that reconstruct the grid model and transform dispersed information into actionable knowledge for distribution grid operation.
Distribution System Operators (DSOs) manage information from multiple systems: GIS, SCADA, AMI, OMS, field measurements, incidents and grid models that are often partial or contain inconsistencies. This heterogeneity makes it difficult to obtain a single, reliable view of the state of the grid. The approach presented by Ariadna in the CIRED Workshop 2026 paper demonstrates that the combination of data governance, advanced algorithms and electrical expertise makes it possible to automatically complement, validate and correct the grid model, even when the available topology is incomplete.
Once structured and enriched, data enables high-value use cases for electrical utilities – DSOs, such as:
- Automatic assignment of Secondary Substation, feeder and phase.
- Immediate identification of customers affected by an incident.
- Support for planned switching operations on the grid.
- Monitoring of technical losses by feeder and phase.
- Detection of electrical losses.
- Identification of behind-the-meter photovoltaic generation.
- Detection of end-of-line overvoltages not visible from the Secondary Substation.
- Automatic phase balancing to improve quality of supply.
Grid digitalisation is not simply about collecting more information. Its real value emerges when data is transformed into reliable information for operations.
A data governance strategy makes it possible to eliminate inconsistencies, connect information from different systems and maintain an up-to-date grid model. On this basis, more accurate grid twins can be built, investment planning can be improved and operational decision-making can be accelerated. The work presented by Ariadna at CIRED shows that even grids with partial information can achieve high levels of observability through the combined use of measurement data, electrical models and validation algorithms.
Electricity distribution networks face stress from electrification, distributed generation and new demand, while low-voltage (LV) grids still suffer from limited observability. To address this, Ariadna and DSOs in Spain and Portugal have implemented an IoT and analytics platform that merges data from secondary substations, LV feeders and smart meters to enable new operational use cases.
DSOs deploy multi-vendor sensors in secondary substations: transformer and feeder devices measuring currents, voltages, power quality and environmental conditions. These measurements, together with smart-meter data (energy, voltage and events) and GIS information, are ingested by the hardware-agnostic IoT platform, which supports multiple manufacturers and communication protocols. Across Spain and Portugal this architecture already covers more than 50,000 secondary substations and 300,000 feeder sensors, correlating data from over five million smart meters. On top of this unified layer, a Grid Twin (GTA) builds an LV digital twin that computes power flows, calculates power losses and voltages, providing the basis for supervision, planning and hosting-capacity analysis.
The first successful key use case is LV topology identification. Connectivity and phase allocation often diverge from GIS records. By correlating secondary-substation sensors and smart-meter profiles over time, GTA infers the real feeder–meter–phase relationships. Deployments in Spain and Portugal show that the topology algorithm reaches 99.9% precision and 95% recall, enabling automatic GIS correction and a trustworthy digital twin.
The second use case is non-technical-loss detection and tampered smart-meter identification. In Spain, UFD’s deployment, documented in an 2025 E.DSO success case (https://www.edsoforsmartgrids.eu/success_cases/ai-powered-fraud-detection/), uses Ariadna’s platform and AI models to compare energy balances between LV feeders and aggregated smart-meter data, enriched with meter events. The system monitors more than three million meters, triggers weekly targeted inspections and achieves around 90% reliability in identifying tampered meters and illegal connections.
The third use case is FIREX, the Fire and Flood Prevention Algorithm. Using the same sensors, FIREX evaluates loading and temperature per feeder, and trends inside each substation. Pattern-recognition models detect incipient overheating and water ingress, generating early alarms that allow DSOs to intervene before fires or floods. Deployments report more than 100 early detections and significant avoided damage, showing how safety and asset-health use cases share the same IoT infrastructure.
In Portugal, E-REDES’ Gridwise project illustrates the impact of this approach at scale. Gridwise, built on this architecture, provides real-time LV monitoring, automatic topology mapping, loss analysis, fraud mitigation and power-quality reporting from sensored secondary substations. The E.DSO Gridwise 2024 success case quantifies regulatory PQ-monitoring savings of about €1.5k per substation (around €250k/year for the first 168 units) and an average reduction of 20 minutes in fault-resolution time, alongside improvements in technical service-quality indicators.
Together, the experiences of UFD in Spain and E-REDES in Portugal show that dense LV grid sensoring with an open IoT platform, a grid digital twin powered with AI analytics improves supervision and control, accelerates fault management, supports loss reduction and delivers a clear, replicable business case.
