The transition to climate-neutral cities requires not only strong policy ambition but also robust, testable, and scalable technological solutions.
Artificial intelligence (AI) is increasingly recognised as a critical enabler in this transformation, supporting smarter energy use, more efficient buildings, and data-driven urban planning.
Yet a key challenge persists: how to validate AI solutions in complex real-world environments before deploying them at scale. The EnergyGuard project addresses this gap through the development of Testing and Experimentation Facilities (TEFs), which provide controlled but realistic environments for testing and validating AI applications.
There are several pilot cases within EnergyGuard TEFs, one of them - the Riga pilot, developed on the data provided by Riga Energy Agency in cooperation with the State Construction Control Bureau of Latvia. Learn more about it here.
This pilot focuses on improving energy efficiency in multi-apartment residential buildings, a sector with significant potential for reducing CO₂ emissions. It demonstrates how AI can be applied in a structured and trustworthy way to support decision-making in urban energy systems.
At the core of the Riga TEF pilot is a Digital Twin of the city’s residential multi-apartment building stock, built on the MATRYCS framework (an EU-funded platform for building energy data analytics).
This Digital Twin integrates multiple data sources into a unified environment, including GEORIGA 3D city models, cadastral data, infrastructure layers, real-time weather and solar potential information, as well as heating system telemetry from more than 4,500 buildings, complemented by historical energy audit data where available.
By combining these datasets, the platform creates a high-resolution, dynamic representation of the urban environment that can be used to simulate and analyse energy performance under different scenarios.
The TEF provides a coherent set of services that together form a complete data-to-decision pipeline.
The first of these is the EPC Dataset Service, which offers harmonised Energy Performance Certificate data stored in the EnergyGuard Data Lake. This dataset includes key building attributes such as geometry, construction year, energy class, heating type, and calculated energy consumption.
By ensuring a high level of metadata standardisation and completeness, the service enables AI developers to train and validate models using reliable and interoperable data.
Building on this foundation, the Digital Twin service provides a geospatial representation of Riga’s multi-apartment buildings enriched with energy-related information.
This allows users to visualise building performance, explore spatial relationships, and integrate data into external applications.
The system is designed to maintain high data consistency and ensure timely updates, thus reflecting real changes in building data with minimal delay.
The third component is the AI-driven Energy Efficiency Decision Support System (EFI-DSS), which translates data into actionable insights.
This tool predicts the impact of renovation measures, generates energy-saving scenarios, and supports both policy planning and investment decisions. By comparing its outputs with available real energy audit data, the system is continuously validated and improved.
It is also designed with a strong focus on usability and explainability, ensuring that decision-makers understand the rationale behind the recommendations provided.
What makes the Riga pilot valuable is its end-to-end validation approach. Instead of focusing on isolated components, the TEF demonstrates how data flows from raw inputs through structured datasets and Digital Twin environments to AI-driven decision-making tools.
This integrated approach ensures that solutions are not only technically sound but also practical and applicable in real urban contexts. In addition, the pilot incorporates mechanisms to assess the trustworthiness of AI systems, including aspects such as accuracy, transparency, and effectiveness, which are essential for adoption in the public sector.
The implications of this work extend beyond individual buildings. By enabling predictive modelling and scenario analysis, the TEF supports more strategic urban planning and more efficient allocation of resources.
Municipalities can use the platform to prioritise renovation projects, optimise funding decisions, and develop evidence-based policies that contribute to long-term emissions reduction.
The ability to simulate renovation scenarios for a large number of buildings, including estimates of energy savings, investment needs, and payback periods, provides a powerful tool for accelerating the energy transition in cities.
Equally important is the role of the TEF as a platform for innovation and collaboration. The Riga pilot is designed for external users, including start-ups, small and medium-sized enterprises, and research organisations.
Through a sandbox environment, developers can test their AI solutions using real datasets and Digital Twin services, reducing the risks associated with market entry and scaling. Public authorities, in turn, benefit from access to innovative tools that can enhance their capacity for planning and decision-making.
The Riga Energy Agency (REA) invites organisations working on AI-based solutions for energy efficiency, building management systems, urban planning tools, and broader smart city applications to engage with the EnergyGuard TEF; REA is open for any further cooperation solutions.
To take a deep dive into how EnergyGuard is testing and validating AI solutions for the energy sector, please visit https://energy-guard.eu
This article was produced by Tālis Linkaits, Project Manager at Riga Energy Agency
- Piesakieties sistēmā, lai varētu publicēt komentārus