SMARTGREENS 2026 Abstracts

Area 1 - Energy Engineering

Full Papers

Paper Nr:	24
Title:	An AD-Hoc Matheuristic to Solve the Operation Problem in Energy Communities
Authors:	Javier A. Gamisel-Muzas, Àngel D. Téllez-Macías, Sara Hatami, Bruno Domenech and Antonin Ponsich
Abstract:	In the current context of energy transition, Energy Communities (EC) emerge as a novel energy framework enabling users to generate, store, share and sell their electricity, in general through renewable energies. Once defined the installation of equipment configuration and sizing, the operation problem of an EC consists in determining the hourly energy flows between prosumers and the national grid, as well as optimal batteries’ state-of-charge. This work introduces an ad-hoc matheuristic for the solution of this highly combinatorial optimization problem, having either an economic or a social objective function. This technique is based on three stages accounting for a statistical analysis, battery assessment and cooperation among prosumers, to heuristically determine the trading role of each user at each hour. Then, a Linear Programming is solved to determine optimal flows and battery states. The computational experiments presented here prove that the proposed technique is able to obtain optimal or sub-optimal solutions, while drastically reducing CPU times with respect to using Mathematical Programming solvers.
Download

Paper Nr:	47
Title:	Curve Fitting-Based Modelling of Electric Vehicle Charging Power Profiles
Authors:	Francesco Calise, Francesco Liberato Cappiello, Luca Cimmino, Massimo Dentice d'Accadia and Maria Vicidomini
Abstract:	The aim of this study is to analyse the charging process of electric vehicles and identify the main parameters that influence it. Based on the comparison with several experimental studies, three representative charging curves were obtained through fitting in MATLAB, considered for different initial state of charge values. Subsequently, through a linear interpolation, a complete model was developed to describe the dynamic variation of the charging power, so that it can be applied in different scenarios. The accuracy of the model was evaluated by comparing the simulated charging times with those of ten real charging sessions. The result showed a mean absolute error of 4.3% and a mean relative error of 0.09%, indicating good consistency with real data. The proposed model can therefore be employed within charging infrastructure simulations to support proper sizing.
Download

Short Papers

Paper Nr:	33
Title:	Unveiling the Potential of Urban Building Energy with 5R3C Thermal Modeling in Energy Communities
Authors:	Antoine Dumont, Chadi Mahfoud, Véronique Feldheim, Paul Lybaert and Sesil Koutra
Abstract:	This paper reviews the role of grey-box Resistance–Capacitance (RC) thermal models within Urban Building Energy Modeling (UBEM) to support Energy Communities (ECs) and building stock decarbonization. Given the growing need for scalable and computationally efficient forecasting tools, RC models are examined as reduced-order approaches that balance physical interpretability and simulation performance. The study analyzes how architectural and operational parameters, including compactness ratio, building age, typology, window-to-wall ratio, and occupancy behavior, influence thermal resistances, capacitances, and dynamic heat demand. The findings highlight the suitability of 5R3C models for district-scale applications, renovation scenario analysis, and renewable integration, while identifying challenges related to parameter calibration, behavioral uncertainty, and large-scale validation.
Download

Paper Nr:	44
Title:	EKINDAR, the New Cooperative Energy Community in Spain
Authors:	Xabat Oregi, Markel Arbulu and Hodei Arzak
Abstract:	The EKINDAR project represents a pioneering medium-scale cooperative energy community in the Basque Country (Spain), designed to democratize access to renewable electricity through shared photovoltaic generation. Developed under the Ekiola initiative, the project enables more than 500 households in Azpeitia to participate collectively in the production and consumption of solar energy without requiring modifications to their domestic electrical systems. EKINDAR integrates technical, economic, legal, and social dimensions, including a detailed membership assessment process, a 1.2 MW photovoltaic installation, and an innovative land-use approach combining energy generation with shrub cultivation. Despite administrative challenges and a contentious legal appeal, the project was ultimately validated by regional and national courts, ensuring its continuity. Since commissioning, EKINDAR has demonstrated the potential of community-driven energy models to reduce electricity costs and strengthen local engagement, while also revealing operational challenges such as the mismatch between generation and consumption patterns and the absence of storage solutions for electric vehicle charging. Overall, EKINDAR illustrates the feasibility and replicability of cooperative photovoltaic initiatives as a key instrument for advancing the European energy transition and fostering socially inclusive renewable energy governance.
Download

Area 2 - Smart and Digital Services

Full Papers

Paper Nr:	32
Title:	Enhancing Collective Self-Consumption through Water Storage Heater Flexibility
Authors:	Pierre-Yves Massé, Maylis Duru, Benoit Couraud, Haicheng Ling, Solal Bizeul, Hugo Roussel, Cléa Verdot, Mariane Vittoz, Estefania Alvarez, Merlinda Andoni, Yann Rozier, Sonam Norbu, David Flynn, Erwin Franquet and Thibault Rihet
Abstract:	While Renewable Energy Communities (RECs) and Collective Self-Consumption (CSC) schemes have emerged as promising tools to accelerate renewable energy adoption and support the net-zero transition, their full potential can only be realised when complemented by demand-side flexibility that aligns consumption with renewable generation. Water storage heaters can function as distributed thermal storage, absorbing excess renewable energy at the community level. This work quantifies both the benefits of water storage heaters flexibility for energy consumers in a CSC community in France (such as energy bill reduction, increase of self-consumption), and the challenges related to the implementation and user acceptance. At the first stage, an annual simulation analysis is performed on a community of 41 households and a large solar PV plant, comparing a scenario without a CSC community, a scenario with a standard CSC community, and a scenario with a CSC community with flexibility from water storage heaters, which showed that an average benefit of 70C/year per household can be achieved due to flexibility and an increase of 6% and 22% of solar PV community self-consumption and self-production respectively. In the second stage, we present the results of the real-world deployment in the community, analysing its technical performance and user reception, and examine the main factors shaping user engagement and satisfaction.
Download

Short Papers

Paper Nr:	15
Title:	A Proposed AI-Based Digital Assistant Architecture for ICD-10-AM Code Identification in Romania’s Healthcare System
Authors:	Mihai Adrian Lungu and Razvan Daniel Zota
Abstract:	This research proposes the design and architecture of a system based on artificial intelligence (AI) that helps Romanian doctors convert their natural language diagnoses into ICD-10-AM codes. In Romania the existing administrative burden and potential for mistakes inherent to coding increases because the country lacks computer-assisted coding tools for diagnostic coding. These tools are essential in clinical record keeping, in public health analysis and in payment for hospital care through a system known as Diagnosis-Related Groups (DRG). Using an advanced form of computer software, the proposed system uses complex processes to automatically suggest ICD-10-AM codes which have been previously coded by hand from notes taken by doctors, and can be integrated with any existing hospital computer system. In this system, suggestions from the computer can be validated or changed by a doctor. The system also gathers feedback in a specific way which aids in its learning and in increasing its performance over time. Personal data is treated in such a way that there is no breach of the EU’s GDPR privacy regulations; this is by the application of privacy by design principles. These principles ensure that personal identifiable information is not stored. In particular, the proposed system’s Romanian language capabilities, full coverage of ICD-10-AM codes and learning mechanism based on feedback stand out when compared with other systems around the world and in this region. A future system, still in its conceptual phase, holds considerable promise to improve coding accuracy, thereby also reducing documentation time, increasing the reliability of DRG-based reimbursements, and facilitating the creation of better healthcare statistics, laying the groundwork for a pilot deployment in Romania.
Download

Area 3 - Smart Infrastructures and Smart Buildings

Full Papers

Paper Nr:	29
Title:	A Real-Time IoT System for Monitoring and Analysis of Thermal Sensation and Clothing Adaptation in Smart Indoor Environment
Authors:	Shalini Kuchhal and Divya Lohani
Abstract:	As humans spend a significant amount of time indoors, it is indispensable to maintain indoor thermal conditions to ensure their physical and mental well-being. However, assessing thermal comfort is challenging owing to the dynamic nature of the environment. To address this challenge, the present research proposes the real-time Internet of Things (IoT)-based monitoring system integrated with an interpretable machine-learning framework to predict thermal sensation mean vote (TSMV) and analyse the factors influencing occupants’ TSMV in natural ventilated educational building. This exploratory analysis frame work used Gradient Boosting regression (GBR) as heuristic approach to explore the clothing adaptation trends and Multinomial Logistic Regression (MLR) as an efficient model to classify the thermal sensation categories and assess the relative influence of environmental and behavioural factors. The efficiency of both the models were validated by key performance metrices. Furthermore, Wald test was performed to indicate the significance of indoor temperature, relative humidity, clothing insulation, and air velocity to predict the occupants’ thermal sensation vote in dynamic indoor environments. The research demonstrates the feasibility of low-cost IoT based monitoring system for thermal comfort assessment and interpretable analysis of occupants’ behaviour.
Download

Paper Nr:	46
Title:	Evaluation of Two Photovoltaic Systems Located on the Same University Campus
Authors:	Aarón Ortiz Peña, Andrés Honrubia Escribano and Emilio Gómez Lázaro
Abstract:	Energy consumption within university campuses exhibits highly heterogeneous patterns driven by building functionality, posing a critical challenge for the integration of renewable energy systems under regulatory constraints. This study presents a comparative, multi-scale analysis of two solar photovoltaic installations operating under a zero-export policy within the same campus: an engineering building (CU1), characterised by continuous research activity, and a faculty of education (CU2), defined by a teaching-administrative schedule. The methodology combines monthly, hourly and correlation-based analyses to quantify demand–generation coupling and the operational impact of zero-export limitations. Results show that CU1 achieves sustained utilisation of solar resources and total savings of 23,465.25 C, while CU2, despite better temporal alignment during teaching periods, is strongly constrained by calendar-driven demand fluctuations, reducing its savings to 11,056.05 C. Periods of inactivity (weekends and August) induce a structural decoupling between generation and demand, forcing severe curtailment and significant renewable energy losses, particularly in teaching-oriented buildings. These findings demonstrate that PV performance is fundamentally conditioned by load profiles and regulatory frameworks. Although both systems are economically viable, research-intensive centres provide a more robust return, whereas teaching buildings require demand-side management or storage to mitigate systematic losses.
Download

Short Papers

Paper Nr:	13
Title:	Bridging Present and Future: 5G and Bidirectional Charging for Scalable Urban Mobility Solutions
Authors:	Shangqing Wang, Christopher Lehmann, Shiwei Shen and Frank H. P. Fitzek
Abstract:	This paper presents a fifth-generation (5G)-enabled bidirectional charging testbed that bridges campus experimentation and an urban pilot deployment in Dresden’s Ostra district (2×50 kW chargers, 3 electric vehicles (EVs), private 5G network). We detail the transition from a direct communication setup to a layered architecture integrating an urban mobility and energy platform (IMEP) for data aggregation and a cloud-ready control and optimisation service (COS) for real-time scheduling, addressing urban scalability challenges such as data flow orchestration, protocol integration, and grid interaction. At the current scale, the control traffic could be handled by wired or long-term evolution (LTE) connections; our use of a private 5G network is therefore a strategic choice to validate an architecture intended to scale to scenarios with hundreds of EVs and dozens of chargers coordinated via cloud-based control. We define end-to-end control latency below 15 ms as a key design objective for vehicle-to-grid (V2G) and vehicle-to-building (V2B) operations and identify practical integration constraints in charger interfaces and state-of-charge reporting that the pilot setup must address. The pilot design also reflects the “faraway water” adoption barrier, where stakeholders hesitate to invest in bidirectional charging absent clear incentives and ecosystem maturity. Building on these insights, we identify critical planning gaps (regulatory standardisation, infrastructure investment) and outline pathways toward city-wide deployment across Dresden’s 76 planned mobility hubs, providing urban planners and infrastructure developers with a blueprint and evaluation plan for scalable smart city energy systems. systems.
Download

Paper Nr:	16
Title:	Control System of Building's Lighting Environment Combining Classification and Reinforcement Learning for Occupant-Centric Comfort and Energy Savings
Authors:	Chloé Helain, Nassim Haddam and Dominique Barth
Abstract:	Ongoing research is being performed on the use of Reinforcement Learning methods to control the lighting of a building in order to generate energy savings. However, challenges arise, among them the presence of several occupants with heterogeneous lighting preferences at different times, and that given users’ preferences may evolve. This work presents a multi-agents control system where each agent is a Reinforcement Learning algorithm to adjust the light intensity, combined with a classification algorithm to categorize different occupant behaviors. In this approach, each agent represents a subset of strategies associated to a cluster of environment states which aims to reflect occupant behavior profiles, and each agent learns and adapts its strategies from the interventions of the occupant on the light control. This new control system is evaluated against a single-agent control system.
Download

Area 4 - Energy-Aware Systems and Technologies

Full Papers

Paper Nr:	18
Title:	Two-Stage Photovoltaic Forecasting: Separating Weather Prediction from Plant-Characteristics
Authors:	Philipp Danner and Hermann de Meer
Abstract:	Several energy management applications rely on accurate photovoltaic generation forecasts. Common metrics like mean absolute error or root-mean-square error, omit error-distribution details needed for stochastic optimization. In addition, several approaches use weather forecasts as inputs without analyzing the source of the prediction error. To overcome this gap, we decompose forecasting into a weather forecast model for environmental parameters such as solar irradiance and temperature and a plant characteristic model that captures site-specific parameters like panel orientation, temperature influence, or regular shading. Satellite-based weather observation serves as an intermediate layer. We analyze the error distribution of the high-resolution rapid-refresh numerical weather prediction model that covers the United States as a black-box model for weather forecasting and train an ensemble of neural networks on historical power output data for the plant characteristic model. Results show mean absolute error increases by 11% and 68% for two selected photovoltaic systems when using weather forecasts instead of satellite-based ground-truth weather observations as a perfect forecast. The generalized hyperbolic and Student’s t distributions adequately fit the forecast errors across lead times.
Download

Paper Nr:	20
Title:	Monetizing Curtailed Renewable Energy through Flexible Bitcoin Mining: A Reinforcement Learning Approach with Real Curtailment Data
Authors:	Ioannis T. Thomaidis, Giannis T. Tsoulfas and Thomas K. Dasaklis
Abstract:	The rapid deployment of wind and solar infrastructure across several countries is outpacing the ability of transmission and distribution grids to absorb variable, location-concentrated generation, resulting in systematic curtailment. Flexible Bitcoin (BTC) mining can be a viable solution in this context. We present a Reinforcement Learning (RL) simulation framework to monetize curtailed renewable energy via flexible BTC mining based on real data. The task is modeled as a two-agent Multi-Agent Reinforcement Learning (MARL) problem (solar, wind) trained with Multi-Agent Soft Actor–Critic (MASAC) under Centralized Training, Decentralized Execution (CTDE), using a centralized twin critic and per-agent squashed Gaussian policies. In the full system, all episodes become profitable at Bitcoin prices of C12,000/BTC and above, while lower prices exhibit occasional losses despite positive average performance. Solar and wind sites display distinct thresholds, reflecting differences in curtailment patterns and operational frictions. Overall, the results demonstrate that flexible Bitcoin mining can monetize a significant share of curtailed renewable energy, but that robust, always-profitable operation requires BTC prices well above average break-even levels.
Download

Paper Nr:	39
Title:	Proximal Policy Optimization for Operating a Hybrid Flywheel–Battery Storage System under Real Factory Load Profiles
Authors:	Sarah Schwarz and Stephan Rinderknecht
Abstract:	As manufacturing shifts toward more sustainable practices, hybrid energy storage systems can enhance operational flexibility and facilitate peak-focused energy management. At TU Darmstadt, the ETA Factory is researching a hybrid storage system that pairs an innovative flywheel with a lithium-ion battery. The flywheel features an outer-rotor design with magnetic bearings to minimize mechanical losses, while the battery subsystem relies on NMC pouch cells. Because of their complementary properties, such as the flywheel’s high power capability and durability during cycling and the battery’s high energy density, hybrid operation promotes efficient power delivery and supports grid-oriented objectives. So far, the system has mainly been operated on a test bench using conventional rule-based control strategies, especially the Optimal Power Share approach. Although these methods are robust and easy to implement, they are usually designed for a limited set of objectives and are difficult to expand when additional criteria, such as battery degradation or predictive information, are included. This work therefore explores deep reinforcement learning as an alternative control approach. A Proximal Policy Optimization (PPO) agent is trained in simulation using real load measurements from the ETA Factory and evaluated over several monthly billing periods. The evaluation uses factory load data with a 30-second resolution from April to September 2021. Training is based on scaled data from April to August, with September serving as out-of-sample test data. The most effective setup employs a reward function that includes peak penalty, storage losses, battery degradation, and export penalty. Compared to a no-storage baseline, it achieves a 13.96% reduction in energy costs for April in-sample and a 9.69% reduction for September out-of-sample. Degradation-related terms are treated as auxiliary objectives and are excluded from the reported energy cost metric.
Download

Short Papers

Paper Nr:	21
Title:	A 24-Hour Global Reference Model to Compare Full-Day and Time-Restricted Residential Load Profiles
Authors:	Andres Muggi, Bruno Domenech and Marc Juanpera
Abstract:	In residential electricity-consumption studies, observations are often restricted to specific time windows of the day, which hampers direct comparison with results obtained from full 24-hour profiles. This situation is common in energy communities and systems with distributed photovoltaic generation, where practitioners often focus on operational periods-such as solar hours-yielding results that are not directly comparable with studies based on complete daily profiles. This paper proposes a methodology to consisteny compare complete and time-restricted views of residential consumption using a shared 24-hour global reference model. The reference is built exclusively from full daily profiles and remains fixed across all comparisons. The method combines (i) adaptive segmentation of the daily temporal domain, which identifies two complementary time windows from the data, and (ii) projection of window-based observations onto the same global reference space. A case study using real residential smart-meter data illustrates the approach.
Download

Area 5 - Smart Cities

Short Papers

Paper Nr:	14
Title:	From Pilot to Practice: Business Model Innovation for 5G-Enabled Bidirectional Charging in Smart Cities
Authors:	Shangqing Wang, Laura del Rio Carazo and Frank H. P. Fitzek
Abstract:	This paper presents a barrier-detection business model framework for 5G-enabled bidirectional electric vehicle (EV) charging, grounded in technical pilots and simulation studies in Dresden’s Ostra district. Using real-world operational data and direct stakeholder engagement, we identify the practical value propositions, cost structures, and incentive mechanisms necessary for scalable Vehicle-to-Grid (V2G) and Vehicle-to-Building (V2B) adoption in urban settings. The analysis systematically maps socio-technical and regulatory barriers, including market, infrastructure, and user acceptance challenges, and proposes actionable, stakeholder-driven pathways for smart city integration. Our findings show that advanced communication networks, combined with collaborative business models, can accelerate the transition to sustainable urban mobility and resilient energy systems. The work delivers a transferable framework for evaluating V2G/V2B integration, supporting both local policy development and broader academic discourse on digital energy transitions.
Download

Paper Nr:	35
Title:	Technical Requirements for a Digital Twin PED: A BIPED Case Study
Authors:	Pavel Kogut, Jiri Bouchal and Susie McAleer
Abstract:	Positive Energy Districts (PEDs) and Digital Twins are increasingly seen as important enablers of the clean urban energy transition promoted by the EU Cities Mission. PEDs integrate buildings, their users, and urban systems, such as transportation, to achieve a net-positive annual energy balance. To help cities capture this multi-sectoral connections and design actionable pathways for PED development, the BIPED project has developed a prototype Digital Twin PED platform. This article employs a descriptive method to provide an overview of the platform's components and illustrate its application with several use cases. It concludes by outlining key technical requirements for Digital Twin PEDs. Its four main components are the data catalogue, model coordinator, 3D environment, and dashboards. Together, they provide an integrated solution for supporting PED related measures across a number of use cases. The paper describes how a Digital Twin PED can be used for rooftop solar planning, building occupancy assessment, traffic management, and district heating mapping. In conclusion, a summary of key requirements for a Digital Twin PED are provided.
Download

Paper Nr:	17
Title:	Mapping Smart and Sustainable Mobility: Challenges, Innovations, and Perspectives for the Future in Portuguese Municipalities
Authors:	Pedro Almeida and Daniela Herculano
Abstract:	This study examines the current stage of Smart City development in Portuguese municipalities, with particular emphasis on urban mobility and the role of cycling as a sustainable transport option. Carried out within the BIKiNNOV project, the research maps out main priorities, existing policies, and emerging directions that influence the evolution of urban environments in Portugal. A structured survey was sent to all 308 municipalities, collecting information on strategic planning, use of digital technologies, mobility infrastructure, and attitudes towards sustainable mobility. The questionnaire also investigated how municipalities participate in innovation ecosystems and collaborate with technology centers. The results offer an integrated picture of how local governments are pursuing smart urban development, highlighting both the opportunities and the constraints they encounter in building sustainable and inclusive mobility systems. By shedding light on municipal priorities and difficulties, this work deepens our understanding of the links between local governance, innovation, and public policy in the move towards smarter, more resilient, and human-centered cities, and provides a useful evidence base to guide decision-making and promote cooperation between public authorities, technology providers, and research organizations.
Download

Paper Nr:	26
Title:	Context-Aware Indicators for Data-Driven Smart City Evaluation: Evidence from the CESMOD Project
Authors:	Zuzana Hruška, Janka Marschalková, Leonard Walletzký and Mouzhi Ge
Abstract:	The smart city concept has evolved into a mainstream framework for urban development, however, evaluating the impacts of smart solutions remains a challenge due to cross-domain effects, context dependency, and limitations in data quality and availability. This paper therefore addresses these challenges by presenting the methodology developed within the CESMOD project in the Czech Republic, which aims to establish a robust, data-driven environment for the long-term measurement and comparison of smart city impacts. The methodology proposes a hierarchical and context-sensitive indicator structure aligned with national smart city strategies while prioritizing data availability and interoperability. To demonstrate the approach through selected component-level indicator sets, the paper has shown how strategic objectives can be operationalized into interpretable and scalable indicators across different municipal contexts. The findings contribute to the improvement of policy usability in smart city evaluation frameworks.
Download

Paper Nr:	31
Title:	Reconstructing Smart Cities through Smart Digital Addressing Systems
Authors:	Ammar Aljer, Ziad Salem and Raneem Alainiah
Abstract:	Conventional urban addressing systems rely on names and postal codes that are not globally standardized. Human-centered solutions restrict automation, real-time communication, and interoperability as cities become digitally connected. This manuscript proposes a machine-oriented digital urban addressing system designed for future smart cities. It shows how such a system is implemented through drawing inspiration from the architecture of the Internet Protocol (IP) addressing and the expansion of Internet of Things (IoT) applications. In this approach, streets, buildings, apartments, districts, etc. are uniquely identified by machine-oriented code that can be converted into human-readable formats. This method offers extensibility, enabling smooth communication between digital devices from various suppliers, and interacts with current city-representative platforms like Building Information Modeling and Geographic Information Systems. The limitations of the current urban addressing systems are also discussed along with a framework for a global system that can enable communication between devices that are physically located at the same address. Developing the concept of such an approach allows for a profound transformation in the very concept of the smart city, leading to greater alignment between its physical infrastructure and its digital framework and facilitating the resolution of many problems arising from this disconnection, such as cybersecurity challenges.
Download