Energy, Civil Infrastructure and Urban Systems
16:05 - 17:25 | Meeting Room 5 (Level 0)
Chair: Chev. Prof. Marc Bonello
Dr Salah Eddine Rhaili
Department of Electrical Engineering, Faculty of Engineering
Offshore floating wind deployment in Malta currently lacks long-term in-situ hub-height measurements. Consequently, reanalysis-based screening serves as the baseline for resource assessment. This study, carried out as part of the OCEAN-H2 project, applies a reproducible workflow for a site within Malta’s Exclusive Economic Zone, using 10 years (2015–2024) of hourly ERA5 wind data. Spatial representativeness within the ERA5 grid is tested by comparing centroid-interpolated and 4-cell-averaged wind speeds, showing a negligible mean difference (0.007 m/s) and near-identical distribution fits (Weibull shape k≈1.71). These results support single-point characterisation. At 100 m, the baseline climatology yields a mean wind speed of 6.58 m/s.
Four hub-height extrapolation methods were assessed to project winds to 150 m and 170 m. Two methods were selected based on robustness and physical plausibility: (i) a neutral logarithmic profile as the primary screening estimator; and (ii) a time-varying shear power-law as a conservative sensitivity bound, observed to under-predict high wind speeds. Gross energy production (AEP) over 2015–2024, computed using the IEA 15 MW and 22 MW reference turbine power curves, is estimated at ~392 GWh and ~398 GWh for the IEA-15 MW turbine (150 m and 170 m), and ~540 GWh and ~550 GWh for the IEA-22 MW turbine (150 m and 170 m). These totals correspond to mean capacity factors of ~0.298 and ~0.303 (IEA-15 MW) and ~0.280 and ~0.285 (IEA-22 MW). Method-induced uncertainty typically remained within 3–4%. These results establish a screening baseline and indicate the need for future in-situ measurement campaigns and loss modelling.
Dr Muhammad Sadiq
Department of Electrical Engineering, Faculty of Engineering
The electrification and digitalisation of seaport infrastructure are turning ports into complex energy systems that integrate onshore power supply, renewable generation, and battery energy storage. Coordinating these assets is challenging owing to the variability of vessel schedules, intermittent renewable energy output, and limited grid capacity. This study focuses on the development of a digital twin platform to support operational energy analysis and COâ‚‚-aware decision-making for port microgrids. The digital twin integrates a dynamic microgrid simulation model with an operator-oriented dashboard. This provides operationally representative insights into power flows, battery state of charge, OPS demand, and grid exchange, within defined technical and contractual constraints. At the core of the platform is a unified framework model that includes key port energy components: photovoltaic generation, battery energy storage systems, port base loads, and vessel hoteling demand. It further enables accounting for resultant COâ‚‚ emissions and performance benchmarking against conventional diesel-based port operations through an integrated module. The results show how the digital twin improves situational awareness, supports what-if operational analyses, and enables the coordinated use of local energy resources to reduce grid reliance and emissions. The platform provides a practical foundation for future smart port applications, including energy management, vessel scheduling support, and data-driven optimisation for more sustainable port operations.
Ms Tram Nguyen
Department of Criminology, Faculty for Social Wellbeing
The MUSAC project (Exploring Malta’s Urban Space with AI Classification of Malta’s 3D CloudIsle) aimed to advance AI-driven analysis of high-resolution 3D LiDAR point cloud data to enable scalable, automated, and objective urban spatial analysis for Malta. Building on national investments in airborne laser scanning, MUSAC addressed key limitations of existing GIS practices, which rely heavily on manual interpretation of satellite imagery and aerial photography. MUSAC delivered an AI-based methodology for semantic classification of urban environments and automated extraction of building contours. An end-to-end processing pipeline was developed, encompassing data standardisation, preprocessing, feature extraction, deep learning based semantic segmentation, and post-processing into GIS ready vector outputs. A dual architecture strategy combining 2D projection-based convolutional neural networks and direct 3D point-cloud processing was implemented and validated across dense urban, suburban, and rural Maltese environments. Class imbalance mitigation, adaptive resolution strategies, and systematic validation ensured robustness and generalisability. Beyond the original research scope, the project produced user-friendly desktop applications and open-source tools that allow non-expert users to extract building footprints and classified urban features directly from LiDAR data. Key outputs include executable applications, an open-source code repository with documentation and tutorials, classified point cloud datasets, and shapefile outputs compatible with standard planning workflows. The project directly supports Malta’s national AI Strategy and establishes a sustainable foundation for future development, evidenced by follow-up funding for the MAiPEX project. MUSAC demonstrates how applied AI research can translate into operational tools for urban planning, infrastructure management, and spatial governance, while strengthening national capacity in advanced geospatial analytics.
Dr Walaa Shoeib
Department of Spatial Planning and Infrastructure, Faculty for the Built Environment
Urban transport authorities increasingly rely on data to manage congestion, plan infrastructure, and respond to disruptions. However, traffic data is often fragmented, noisy, and incomplete, limiting its practical value for real-time decision-making, particularly in dense, small-island urban environments such as Malta. This research addresses these challenges through intelligent data fusion and modern traffic modelling methods designed for data-constrained road networks.
The work focuses on improving the accuracy of origin-destination (OD) estimation by integrating heterogeneous data sources, including roadside cameras, anonymised automatic number plate recognition (ANPR) counts, vehicle trajectories, and traditional surveyed traffic measurements. By explicitly combining probabilistic estimation techniques with physical road-network constraints, the proposed framework enables robust inference of vehicle movements even under sparse camera coverage, noisy observations, and intermittent data loss. Initial investigations centre on single- and multi-junction case studies, where ANPR-derived entry-exit matching enhances OD estimation accuracy, with a clear pathway towards scalable, network-wide application.
Beyond measurement, the research incorporates macro- level traffic flow models with integrated predictive methods to support scenario-based analysis of planned and unplanned disruptions, such as roadworks, public events, or accidents. These models allow transport operators to assess likely impacts, explore re-routing strategies, and support evidence-based decision-making in near real time. The resulting outputs also inform the development of tailored key performance indicators (KPIs) specifically designed for small, high-density urban networks, addressing limitations of conventional metrics developed for larger cities.
Overall, this research demonstrates how ANPR-enhanced, uncertainty-aware traffic modelling can provide a scalable, low-cost foundation for more resilient and responsive urban traffic management.
Dr Jeanette Muñoz Abela
Department of Civil and Structural Engineering, Faculty for the Built Environment
Artificial intelligence is increasingly discussed as a transformative tool within civil engineering; however, its actual adoption across Europe remains uneven and cautiously framed. This paper presents the findings of a qualitative questionnaire distributed among national chambers and professional bodies representing civil engineers across Europe, coordinated through the European Council of Civil Engineers. The survey explores current applications of artificial intelligence, perceived barriers to adoption, ethical considerations, and anticipated impacts on education and professional practice.
The results indicate that artificial intelligence is presently used primarily in assistive and low-risk contexts, including document preparation, project planning, early-stage design support, and educational activities. More advanced or autonomous applications, particularly those influencing structural decision-making, remain limited. Ethical concerns are consistently highlighted, especially regarding accountability, data integrity, professional liability, and the transparency of AI-generated outputs. These concerns are framed pragmatically, reflecting the profession’s responsibility for public safety and long-term performance of the built environment.
Regional variations are evident. Respondents from Western and Northern Europe generally report higher levels of awareness and experimentation, while Southern and Eastern European contexts emphasise constraints related to skills, resources, and access to digital infrastructure. These differences suggest an emerging form of AI inequality within the profession, driven more by organisational capacity than geography alone.
Education and continuing professional development are identified as the most appropriate entry points for responsible AI integration, offering opportunities to enhance learning while reinforcing core engineering judgement. Overall, the findings portray a profession in transition rather than crisis, facing the challenge of integrating artificial intelligence in ways that preserve competence, accountability, and societal trust.
Prof. Vincent Buhagiar
Department of Environmental Design, Faculty for the Built Environment
The DCB, the Double C-Block, is based on a unique design that could revolutionise the building industry. Its geometry replaces the hollow core block (HCB) with twin C-shaped concrete elements, bonded with three embedded layers of insulation to produce a readily insulated block rolled out of the production line. Following the conceptual development and rigorous testing of the DCB in test cells, it is now ready for field studies, including capturing human feedback in a real working office environment. This is being done at the Sustainable Living Centre (SLC), a modern office building on UM Campus, Msida. After careful consideration, Block A was identified as the ideal location for installing the DCB as a pre-insulated perimeter wall, facing three key orientations for direct cross comparison. The third floor was chosen from four to enable a direct floor-to-floor comparison of similarly laid-out offices, at the second floor, built with the HCB. Both are sheathed by a floor each, avoiding the influence of exposed floor and ceiling surfaces to the outdoor environment. Thermocouples are being installed in each office space, with additional surface temperature sensors to assess the thermal performance of the walls on different floors. Separately, acoustic tests will also be carried out to compare the different floors and spaces. Questionnaire surveys will collect subjective feedback to compare with data logging and simulations of the walls, when compared to the test cells. The methodology for such monitoring will be presented in the pitch.