Driving Smarter Energy Communities Through Advanced Forecasting
DUTH’s Contribution to HARMONISE
As Europe accelerates its transition toward sustainable and intelligent energy systems, accurate forecasting has become a critical enabler of efficient energy management, flexibility optimization, and renewable energy integration. Within the Horizon Europe project HARMONISE, the Democritus University of Thrace (DUTH) is playing a key role in advancing these capabilities through the development of innovative forecasting solutions for smart energy communities.
Advancing Forecasting Innovation in HARMONISE
At the heart of DUTH’s contribution to HARMONISE are Prof. Stelios Krinidis and PhD Candidate Ioannis Tzitzios, who are actively involved in Work Package 3 (WP3), Task 3.1, focused on the development of the HARMONISE Awareness and Forecasting Toolbox.
Their work centers on the design, implementation, and validation of advanced forecasting tools capable of predicting photovoltaic (PV) generation and electrical load demand in smart energy communities. These forecasting solutions are essential for enabling more efficient operation of energy systems and supporting the project’s broader objectives of flexibility management and renewable energy integration.
The team is developing sophisticated AI-based forecasting algorithms using state-of-the-art machine learning models, including XGBoost, LightGBM, CatBoost, and Hybrid Meta-Models. These models are integrated with real-time data streams provided through the HARMONISE HIP Platform, enabling continuous forecasting and operational support for energy communities.
In addition to model development, the DUTH team is responsible for validating forecasting performance using real-world pilot data, with particular emphasis on the Swiss pilot site. Their work also contributes to predictive analytics tools that support flexibility management and energy optimization across the HARMONISE ecosystem.
Collaboration Across Europe
Now in its second year, HARMONISE has demonstrated the value of strong international collaboration. According to the DUTH team, one of the most impressive aspects of the project has been the close cooperation among partners from different countries, disciplines, and sectors.
By combining cutting-edge research with real-world implementation, HARMONISE creates an environment where innovation can be rapidly tested, validated, and applied. This multidisciplinary collaboration allows academic researchers, industry experts, software developers, and pilot operators to work together toward shared goals, ensuring that technological advancements can deliver tangible benefits to energy communities across Europe.
Tackling Complex Data and Real-Time Challenges
Like many ambitious innovation projects, HARMONISE has faced technical challenges along the way. One of the primary obstacles encountered by the DUTH team has been the integration and synchronization of heterogeneous datasets originating from multiple pilot sites and platforms.
Variations in data quality, temporal resolution, and data availability required extensive preprocessing and validation efforts before the forecasting models could be effectively deployed.
Another significant challenge involved optimizing forecasting performance under real-time operational conditions. To address this, the team developed hybrid approaches that combine physical models with advanced machine learning techniques. This combination has enhanced both the robustness and accuracy of the forecasting tools, allowing them to perform effectively in dynamic and operational environments.
These experiences have also generated valuable lessons for the wider consortium. The team highlights the importance of standardized data formats and continuous communication among technical teams. Strong collaboration between software developers, researchers, and pilot operators has proven essential for improving implementation efficiency and achieving project objectives.
A Major Milestone: Validating the Hybrid Meta-Model
Among the most rewarding moments for the DUTH team has been the successful validation of the Hybrid Meta-Model for photovoltaic forecasting using real pilot data.
Achieving strong forecasting accuracy while integrating the models with live data streams from the HIP Platform represented a significant milestone. It demonstrated not only the technical effectiveness of the forecasting toolbox but also its readiness to support real-world energy management applications.
The project’s collaborative atmosphere has also been a highlight. Regular meetings and technical discussions have fostered a rich exchange of expertise between academia, industry partners, and pilot stakeholders, creating an inspiring environment for innovation and knowledge sharing.
Creating Value for Energy Communities
The impact of DUTH’s work extends beyond the project itself. Advanced forecasting solutions play a crucial role in the development of intelligent and flexible energy systems capable of supporting Europe’s energy transition.
By improving the accuracy and reliability of renewable energy forecasting, the HARMONISE forecasting toolbox helps stakeholders make better-informed decisions, optimize energy use, and proactively manage energy resources. These capabilities contribute to enhanced grid stability, improved integration of renewable energy sources, and reduced operational uncertainty.
For energy communities and pilot sites, the forecasting toolbox provides practical decision-support capabilities that enable more efficient and resilient energy management strategies.
Looking Toward the Future
As HARMONISE progresses into its next phases, the DUTH team is particularly excited about expanding the validation of its forecasting models across additional pilot sites and testing their performance under diverse real-world conditions.
Future research opportunities include the development of adaptive AI systems, real-time optimization approaches, and predictive flexibility management solutions for smart grids and Positive Energy Districts (PEDs). These advancements have the potential to further strengthen the role of intelligent forecasting in creating more sustainable, resilient, and efficient energy communities.
Strengthening European Innovation Through Collaboration
For the DUTH team, HARMONISE represents more than a research project—it is a demonstration of how collaborative European innovation can accelerate progress toward sustainable energy systems.
Participation in the project has provided valuable technical experience while strengthening international scientific collaboration. By bringing together expertise from across Europe, HARMONISE is helping to transform advanced research into practical tools that support smarter and more sustainable energy communities.
“Through HARMONISE, we are transforming advanced forecasting research into practical tools that support smarter and more sustainable energy communities across Europe.”