Resources
D4.2 CRETE VALLEY Digital Twin, data-driven services & apps (second wave)
This document presents the second wave of data-driven applications and services aimed at supporting the System-of-systems Digital Twin (DT) as a digital representative of CRETE VALLEY’s assets, primarily focused on the island of Crete, Greece. The integrated solution unites multiple advanced modules and tools into a cohesive architecture that supports grid operators, energy suppliers, and consumers in optimizing energy flow, improving operational efficiency, and promoting energy autonomy.
Conference proceeding: Enhancing Engine Mount Design through Topology Optimization for Additive Manufacturing
Additive manufacturing and topology optimization are enabling the development of lightweight, high-performance mechanical components with complex geometries. This study investigates an automotive engine mount bracket using 3D scanning, finite element analysis, and topology optimization, comparing the mechanical performance of topology-optimized, gyroid, and auxetic designs.
Conference proceeding: A meta-learning framework for short-term wind power forecasting with SCADA and weather data
Forecasting wind power accurately remains a critical challenge in integrating renewable energy into modern power systems. In response to this challenge, this paper presents a hybrid ensemble learning framework for wind power point forecasting, combining multiple Machine Learning regressors through a meta-learning strategy. The approach leverages historical SCADA measurements, weather forecasted data and physically derived theoretical power curves to enrich the feature space.
Scientific publication: Less Is More: Data-Driven Day-Ahead Electricity Price Forecasting with Short Training Windows
Accurate electricity price forecasting in volatile Day-Ahead Markets requires models that adapt quickly to changing conditions. This publication proposes a short-window learning approach (7–90 days) and compares deep learning with boosting methods across European markets. Results show that LightGBM delivers superior accuracy and robustness, particularly in capturing seasonal patterns and peak price events, supporting efficient and reliable forecasting under data-scarce conditions.
Conference proceeding: Modular Load Estimation Framework for Low-Observability Distribution Networks: Crete Case Study
Limited data availability remains a key challenge for monitoring and managing distribution networks, especially where smart metering is not widely deployed. This paper proposes a modular load estimation framework designed for low-observability systems, using feeder-level measurements and adaptable estimation strategies. Validated through a real case study in Crete, the approach demonstrates that accurate load and voltage estimates can be achieved even under severe data constraints.
Conference proceeding: SARGON2 Ontology: Updating SARGON Ontology for Semantic Data Sharing in the Energy Domain
The growing complexity of modern energy systems requires effective integration of heterogeneous data across multiple domains. Building on Semantic Web technologies, this publication introduces SARGON2, an extended ontology designed to improve data interoperability in the energy sector. By incorporating renewable energy sources, weather information and power system components, the framework enables more consistent and machine-readable data exchange, supporting the development of advanced smart energy applications.
Scientific publication: Application of Measure–Correlate–Predict (MCP) Methodology for Long-Term Evaluation of Wind Potential and Energy Production
Accurate long-term assessment of wind resources is essential for reliable wind farm development, particularly in complex mountainous terrains. This publication presents an integrated approach combining short-term ground measurements with long-term ERA5 data through an advanced measure–correlate–predict method. By incorporating uncertainty analysis and terrain effects, the framework provides realistic estimates of annual energy production and exceedance probabilities, supporting more robust and bankable wind energy projections.
Conference proceeding: Deployment of Modular Renewable Energy Sources and Energy Storage Schemes in a Renewable Energy Valley
Community energy initiatives have shown strong potential at local level, but their site-specific nature often limits wider scalability. Building on real-world demonstrations in Crete, this paper introduces a modular framework for Renewable Energy Valleys, combining residential demand with different sectoral applications such as tourism, rural heating and transport. By optimally integrating solar, wind, battery storage and hydrogen systems across varying renewable energy targets, the approach identifies flexible and efficient configurations that enhance reliability and reduce external energy dependence. The results highlight how such modular solutions can support the scalable development of sustainable energy systems across diverse regions.
Scientific publication: The Evolution of Windmill Design: From Lasithi Plateau Pumping Windmills to Electricity Production
Traditional windmills of Lasithi Plateau have long supported local water supply through their distinctive fabric-sail rotor systems, designed to operate under varying wind conditions. Building on this heritage, the study examines the aerodynamic performance and structural behaviour of such systems, with particular attention to sail geometry, supporting structures, and passive deformation mechanisms. The results highlight how flexible sails can naturally regulate loads and enhance resilience in high winds, offering valuable insights for the design of modern, small-scale wind turbines inspired by low-cost and robust traditional solutions.
Scientific publication: Transfer learning techniques on temporal fusion transformers for short-term building load forecasting under limited data conditions
The evolution of energy systems toward smart grids is increasing the availability of building-level data, yet many buildings still face limited historical records, hindering accurate short-term load forecasting. Transfer learning offers a promising solution by enabling models trained on data-rich buildings to be adapted to data-scarce environments. By testing several forecasting models, including the advanced Temporal Fusion Transformer, this publication shows that this approach consistently improves prediction accuracy, with notable reductions in error and strong performance across different building contexts.
D3.2 – Energy Data Space compliant tools for sovereign data sharing, and extended interoperability in REVs (2nd technology release)
This document presents the second iteration of the CRETE VALLEY digital backbone, marking the transition from a reference architecture to the concrete technical implementation of the Energy Data Space. It describes the functional deployment and integration of key components that enable sovereign data sharing and advanced interoperability within and across Renewable Energy Valley Living Labs in Crete. Building on the foundations defined in the first release, this update demonstrates the operational maturity of the Energy Data Space, including secure data exchange, semantic interoperability and a decentralised peer‑to‑peer marketplace.
D8.2 – Second release of the dissemination & communication plan, & progress report
This document provides an updated Communication and Dissemination Plan for CRETE VALLEY, reflecting progress achieved during the first half of the project and outlining priorities for the next phase. Building on the initial strategy, it updates key messages, performance indicators and engagement results across communication channels. The document reviews dissemination and communication activities implemented from month 1 to month 28, assesses their effectiveness through monitoring and evaluation, and presents a structured plan for future actions aimed at supporting exploitation, visibility and long‑term project impact.