This paper proposes a deep learning based pipeline for generating semantic meshes and relationship graphs from intensity only laser point clouds of hydropower plants. The approach segments point clouds into object-level instances using supervised models and derives spatial relationships such as proximity and containment between components. These relationships are then refined into a structured graph, providing a lightweight yet informative representation suitable for digital twin applications. By replacing traditional rule-based methods with learning-based segmentation, the pipeline achieves higher accuracy and produces more reliable relationship graphs, enabling scalable and automated modeling of complex dam infrastructure.
Although pre-demolition audits (PDAs) provide early information on building materials, they are underutilized partly because their data require domain expertise to access and interpret. This raises the question: can natural-language querying enable non-experts to retrieve information from PDA datasets? We propose a knowledge-graph-based Retrieval-Augmented Generation (RAG) framework for natural-language querying of PDAs, applied to a case study of 40 French audits. Evaluated against a semantic RAG baseline across 21 queries, Graph-RAG achieves a success rate of 90% compared to 57% using vector-based retrieval, though at 2.34× higher cost, demonstrating its potential to making PDA content accessible to non-expert stakeholders.
The bridge design process is a complex and time-consuming task for bridge engineers. To reduce repetitive tasks, automating the application of guidelines and standards is desirable. Currently, automation attempts are mostly based on table-based calculations. This work proposes a new approach to automated, code-compliant structural design: an exemplary automation process for dimensioning the bridge cap. In this process, an ontology describing bridge components and boundary conditions is created. German road bridge design rules are implemented using the Semantic Web Rule Language (SWRL). The aim is to enable a fully automatic geometric dimensioning of bridge caps based on several boundary conditions.
This study investigates where and why material identification and digital representation fail to persist across reuse life cycles. Digital material tracking in circular construction is commonly organized through separate data pipelines for logistics and design, resulting in discontinuities as building elements move between projects. A unified analytical framework is proposed and used to define six recurring classes of breakpoints in data continuity. These breakpoints are empirically identified and analyzed across eight construction reuse case studies. The findings establish a systematic breakpoint typology and provide recommendations for continuity-oriented data infrastructures that support traceable material reuse in construction.
Public administration buildings generate vast operational data, which is rarely used for environmental purposes. GreenBIMxels, a research project led by the Polytechnic University of Catalonia, explores how Data Spaces support environmental assessments in public facilities.
This paper defines environmental services for BIM models within a Data Space named UPCxels. The study addresses the interests of public administrations and private-sector stakeholders, examining how Data Spaces can respond to their needs. Three use cases are designed by combining participatory workshops and a questionnaire survey. The use cases include energy-efficiency, water-consumption optimization and carbon-footprint assessment, each operationalized through KPI sets and BIM/IFC-based properties.
Smart Mobile Factories (SMFs) bring prefabrication to the construction site, enabling just‑in‑time delivery, shorter haulage distances and lower carbon emissions. Although integrated digital twins and modular factory designs have demonstrated the technical feasibility of SMFs in construction, their value‑creation mechanisms for operators and clients remain unclear. Adapting the Business‑Model Canvas to SMFs, we identify four emerging archetypes. Two case vignettes illustrate the transition from a low‑tech, asset‑centric to a high‑tech, data‑centric service model. Finally, we propose a mapping that relates organizational digital maturity to suitable SMF business-model archetypes, which allows to indicate a possible path toward profitable and sustainable adoption.
Enabling predictive maintenance (PM) of road infrastructure by understanding condition developments requires accounting for internal and external influences. However, the corresponding data are highly heterogeneous and distributed, limiting comprehensive large-scale analysis. To bridge this gap, an existing Digital Twin (DT) approach is enhanced with environmental context data. Knowledge graphs enable scalable querying across heterogeneous subsystems, while a federated database addresses data distribution. The framework is validated using real-world data from Germany. Linking historical climate records with bridge condition development demonstrates comprehensive querying. By enabling scalable cross-system data association, the proposed DT framework enhances the practical applicability of existing PM methods.
As one of the primary information carriers throughout the bridge lifecycle, drawings contain rich semantic information. Traditional file-based management does not support efficient organization and retrieval. In this work, an ontology-based semantic management method for bridge drawings is proposed, leveraging information extracted by artificial intelligence (AI) models. After processing, AI outputs are transformed into structured semantic entities and relationships defined in the ontology. The resulting knowledge graph enables semantic search and helps engineers trace how structural components appear and evolve across drawings. It enhances the accessibility and reusability of bridge drawings and provides a practical foundation for intelligent asset management.
Remote work has increased the need to understand how indoor spatial design affects occupants’ cognitive performance and physiological states. This study proposes a Human-sensored Architecture framework integrating Mixed Reality (HoloLens 2) for real-time spatial manipulation and EEG (Emotiv EPOC+) for physiological measurement. Following a network analysis of 124 publications, experiments with 30 participants revealed that red walls enhanced short-term executive function through stress-induced arousal, whereas preferred, low-stress settings improved long-term memory and productivity. These findings demonstrate the value of human-centered spatial design tailored to specific cognitive tasks and provide a foundation for personalized, adaptive architecture.
Title-blocks in engineering drawings contain essential metadata, but automated interpretation remains difficult because of historical scans and heterogeneous layouts. This paper presents a vision-language model (VLM)-based workflow for title-block detection and structured information extraction using the open-source Qwen2.5-VL backbone. The workflow combines promptable visual grounding on full drawings with schema-constrained extraction from cropped title-block images. The study focuses on three representative fields: scale, drawing name, and route. Results show promising performance for both tasks and indicate that open-source VLMs can provide a simplified alternative to conventional multi-stage pipelines for architecture, engineering, and construction (AEC) document analysis.
