Abstracts Track 2026

Nr:	23
Title:	Energy-Flexible Motion for Reciprocating Mechanisms
Authors:	Foeke Vanbecelaere, Théo Tuerlinckx and Zhenmin Tao
Abstract:	Manufacturing companies face fluctuating energy availability due to the mix of renewable sources, on-site production, and grid supply. To address this, machines need to operate in an energy-aware manner, for example, by adjusting their speed in real-time. Reciprocating mechanisms, commonly used in industry, typically use point-to-point motion profiles optimized for fixed speed and load. When the central energy controller requires speed adjustments, these profiles become sub-optimal, increasing the energy consumption. This work aims to achieve energy-optimal motion profiles across varying production speeds and integrate them into a central energy controller (CEC) managing multiple machine variants. A plant with five parallel machines, each having a different load, is modeled as proof of concept. The motion profiles are parameterized using Chebyshev polynomials, enabling numerically efficient optimization. The optimal control problem, which involves finding the energy-optimal profiles while meeting productivity constraints, is solved off-line. The results are stored for the decision-making of the CEC. Simulations demonstrate that the CEC successfully reduces grid usage while the machines adapt their speed to maximize on-site energy usage. A benchmark is performed using standard motion profiles that meet productivity constraints but are not energy-optimal. Using optimized profiles without the CEC reduces energy consumption by 30 % for the same number of products produced, but still requires grid usage. Activating the CEC effectively eliminates grid usage while still maintaining the same number of products. The simulations demonstrate the potential of energy-flexible motion profiles in a simplified plant model. Future work will extend CEC strategies to include not only production speed adaptation, but also production rescheduling and on-site energy buffers. This approach supports energy-efficient and sustainable manufacturing without any hardware modifications.

Nr:	37
Title:	Assessment of Non-Intrusive Load Monitoring Algorithms on Edge Platforms for Smart Grid Applications
Authors:	Gustavo Gabriel García Acevedo, Javier E. Sierra, Alejandro Sallyth Guerrero Hernandez and Suad Salcedo
Abstract:	The transition toward smart grids demands energy monitoring solutions capable of operating close to the data source, with low latency, reduced cloud dependence, and stronger privacy guarantees. In this context, Non-Intrusive Load Monitoring (NILM) has emerged as a promising approach for appliance-level consumption identification from aggregated electrical signals. However, the practical feasibility of deploying NILM algorithms on edge devices with constrained computational resources remains an open challenge. This work presents an experimental assessment of the computational and functional performance of two representative NILM approaches: an event detection model based on DeepDFML-NILM and an energy disaggregation model based on Seq2Point-CNN. Both models were trained on a high-performance computer and subsequently evaluated on three platforms: a reference PC, an NVIDIA Jetson Nano, and an NVIDIA Jetson Orin Nano. Public and experimental datasets were employed, together with a reproducible execution environment supported by containers and virtual environments. The evaluation considered inference time, CPU, RAM, and GPU usage, as well as detection, classification, and disaggregation metrics under near-real-time operating conditions. The results show that edge devices can preserve competitive predictive performance with respect to the reference PC in both detection and disaggregation tasks. Nevertheless, memory constraints and processing latency significantly affect responsiveness under continuous inference scenarios, particularly on the Jetson Nano. The findings suggest that edge-based NILM constitutes a technically viable alternative for decentralized energy monitoring and local deferred analysis. However, further model and runtime optimization is still required to support strictly real-time operation in smart electrical systems.

Nr:	40
Title:	Data-Driven Analysis of Urban Bus Delays
Authors:	Martin Žnidaršič and Gašper Kreft
Abstract:	Within the scope of efforts to promote a more sustainable, smart and user-friendly transportation in one of Slovenian cities, we studied the bus delays and the factors that affect them. In general, the public transport reliability factors were a subject of several existing studies, in various contexts and with diverse sets of variables observed. Our work, however, is focused on a specific local context for which such studies were not done yet. An in depth study of this topic was motivated by an assessment of potential changes of the bus lines in the city and by an interesting observation that the anecdotal reports and survey responses of common lengthy delays did not match the delay measurements in real-world, as the measured delays were in most cases relatively small. GPS data and time measurements were used for comparison of real-world situation and the timetable schedule, while additional variables that we considered include weather data, ride features (bus stops, length), time and date features and, for a subset of data also the bus occupancy. Performance of our machine-learned predictive models (e.g., coefficient of determination above 0.6) on testing data indicates some predictive potential of the observed variables. According to our random-forest feature importance analysis, the time of day, ride length from the start and some specific bus lines are the most relevant predictive features. When incorporating occupancy data for a subset of routes, the occupancy emerged to be the most relevant feature by far. Occupancy is also the most positively correlated individual variable with bus delays, which to some extent explains the frequency of anecdotal reports of lengthy delays and suggests corresponding adaptations of the delay measurement approach.

Nr:	42
Title:	Enhancing Data Literacy in Smart Homes through User-Centric Training Concepts
Authors:	Hendrik van der Valk and Annika Hesse
Abstract:	We address the gap between complex, data-driven smart home systems and limited user capabilities. While technologies offer automation and decision support, users often lack the data literacy necessary to manage them. We investigate how user-centered training programs can be designed to promote data literacy in smart home environments. We adopt a participatory and design-oriented research approach. We conduct a comprehensive literature review and guided interviews to ascertain user challenges, needs, and preferences. We use co-design workshops and design thinking to develop training materials, such as online courses and self-guided exercises. We engage real-world users to test the effectiveness of the materials. Current observations indicate that users have significant concerns about data breaches and are largely unaware of how their data is collected and used. In everyday contexts, non-expert users struggle to interpret and reflect on system-generated data. Limited mental models and difficulties in understanding automated decision-making processes present substantial barriers to technology acceptance. Existing research has predominantly focused on data literacy within educational and organizational frameworks. There is a notable scarcity of studies examining the smart home sector. Further information is required regarding specific measures that can be implemented to protect and empower users within their private living spaces. By fostering data literacy, we intend to improve users’ ability to evaluate the processes and make informed, critical decisions. This enhances transparency, trust, and perceived control. These measures support the sustainable and confident use of smart home technologies. The work’s originality lies in its user-centricity, placing the citizen at the heart of the design process. It provides a novel contribution to an under-researched field by bridging the gap between sophisticated data aggregation and practical user empowerment in smart homes.

Nr:	43
Title:	Best Practices for User-Centric Smart Home Applications
Authors:	Annika Hesse and Hendrik van der Valk
Abstract:	Our research addresses the development of data-privacy-compliant data ecosystems within smart homes and smart cities. The objective is to identify and consolidate best practices for user-centric applications that allow citizens to monitor data flows transparently. We aim to empower individuals to exercise their fundamental right to informational self-determination in an increasingly automated society. The study utilizes a multi-step methodology. We conduct a Systematic Literature Review and document it via PRISMA standards. A concept-matrix methodology is used to uncover patterns and relationships among existing approaches. An Analytic Hierarchy Process facilitates the development of a weighted evaluation table, culminating in a taxonomy to classify best practices across defined dimensions. We identify a consolidated set of best practices, emphasizing that transparency of data flows, participation mechanisms, and clear governance are critical success factors. We provide a prioritized framework and classification system that highlight the most influential factors for user acceptance and sustainability. Findings indicate that long-term success is not solely dependent on technological performance but requires governance structures that encourage voluntary engagement. A significant challenge remains in resolving the dilemma individuals face between privacy protection and the perceived utility of smart technologies. By fostering trust and acceptance, this work supports the enhancement of civic engagement and quality of life. Technically, it promotes efficiency gains and the development of innovative, socially and environmentally sustainable value-added services. The originality of this research lies in its transition from theoretical best practices to practical application through the design of a practitioner's toolbox. The ultimate value is demonstrated by our commitment to implementing and validating these frameworks within a functional smart home and city ecosystem.

Area 1 - Smart Infrastructures and Smart Buildings

Nr:	41
Title:	BlueBird: Combining Social and Technical Energy Flexibility to Render Buildings Active Nodes in Energy Grids
Authors:	Sonja Klingert, Gerard Laguna Benet and Mona Bielig
Abstract:	The next generation energy system needs flexibility in energy utilization to ensure stability of service delivery and a high utilization of renewable energy sources. This is well known for the power system where volatile renewable energy generation will need adaptation from the grid user side to guarantee that the inflow into the and outflow from the grid are equalized at all times. To a lesser extent, this is also true for heat grids, due to inertia. The EU-Project BlueBird looks at energy flexibility from the point of view of buildings and building-users, aiming at turning buildings into active nodes in energy grids through combining social and technical energy flexibility. Technical energy flexibility means to optimize all controllable assets inside a building to respond to requests from energy markets or energy grids, to either increase or decrease the load on the grid at specific periods of time. Use cases span capitalizing on building energy management systems that operate the building as a whole (that means across the sectors heat and electricity) over asking drivers of electric vehicles to give more charging time for smart-charging the cars to using pumps in water supply storages as means of electrical demand response. In some of these use cases, no interaction with final energy consumers is needed, in others, the mere technology acceptance of affected people is enough, trusting in central control to implement optimization for them without compromising on their comfort or other areas of well-being. In yet another set of cases, the active participation of final energy users is required, for instance entering a deadline for smart-charging their car into a tool or shifting the time of using their washing machine to a high solar irradiation period of time. Both technical and social flexibility, as well as their interaction, depend on a variety of conditions to have a real impact on the grid or the utilization of renewables. The talk will focus on both theoretical and practical expectations and limitations, capitalizing on findings from the heterogenous set of 7 pilots across Europe that implement multiple core use cases affecting a variety of different types of assets.

Area 2 - Energy-Aware Systems and Technologies

Nr:	38
Title:	Novel Energy Networks
Authors:	Francesco Calise, Francesco Liberato Cappiello, Maria Vicidomini, Luca Cimmino and Massimo Dentice d'Accadia
Abstract:	The transition toward a fully decarbonized energy scenario must address the challenge of managing excess renewable production. In this context, smart energy networks are essential to ensure system robustness, resilience, and stability. These networks enable bidirectional flows and integrate innovative solutions for storage and system management. Power to X (PtX) technologies, district heating and cooling networks, renewable energy communities, sector coupling, demand response strategies, the water–energy–food nexus, and advanced storage systems will play a key role in future energy systems. Improving energy efficiency in buildings and transport is another milestone in the decarbonization process. In this framework, the University of Naples Federico II (UNINA) is involved in several projects addressing different aspects of these challenges. Among the largest initiatives, the NEST project develops a network for analysis, simulation, and digital twinning of technologies relevant to the energy transition. Within NEST, UNINA investigates: a) hydrogen systems, electrolyzers, metal hydrides, and fuel cells; b) Power to Methane and Power to Ammonia processes; c) electrical storage based on thermodynamic cycles; d) thermal storage using phase change materials; and e) 4th and 5th generation district heating and cooling networks. These technologies are assessed through detailed dynamic models and co simulation approaches. Various layouts and case studies have been evaluated from energy, environmental, and economic perspectives to determine their future feasibility. The GRETHA project focuses on a novel electrical storage prototype for remote applications. The system uses excess renewable energy to produce hydrogen via electrolyzer, stores it in metal hydrides, and converts it back into electricity when demand exceeds renewable generation. A prototype will soon be tested on a remote site, with plans for future scaling, including airport applications. The project also developed a simulation tool enabling the analysis of different configurations, such as combined hydrogen production through methane reforming and use for electric or fuel cell vehicles. The OPTIMISM project created a comprehensive platform for the design and optimization of energy networks in non residential buildings. It has been applied to office buildings, university facilities, and hospitals, and includes detailed CFD simulations for specific issues. The AGRINEW project investigated energy production from food industry waste. A case study on a cheese factory in Campania demonstrated the use of residual whey in an anaerobic digester producing biomethane to fuel a cogenerator for electricity, heat, and cooling, supported by a large PV field. Biomethane production was further explored in the BIOFEEDSTOCK project, which analyzed multiple configurations and upgrading processes. The group is currently involved in new proposals on decarbonizing hard to abate sectors, developing positive energy districts (PEDs), and advancing storage technologies. Research focuses on CO₂ based closed loop storage systems and high temperature thermal storage within power to heat strategies. For PED development, integrating energy engineering, control science, and social sciences is considered essential.