Pei-Yu Wu


Last Name

Wu

First Name

Pei-Yu

ORCID

0000-0002-2178-5391

Organisational unit

09750 - De Wolf, Catherine / De Wolf, Catherine

Filters

2025 - 20265
Reset filters

Search Results

Publications 1 - 5 of 5
  • Fostering creativity using AI towards a circular economy in architectural engineering design
    Item type: Conference Paper
    Schwarzkopf, Vanessa; Wu, Pei-Yu; Nolte, Tobias; et al. (2025)
    Structures and Architecture ~ Structures and Architecture - REstructure REmaterialize REthink REuse
    Introducing repurposed materials into building design is critical for reducing greenhouse gas emissions and waste generation from the architecture, engineering, and construction sector. However, designing from reclaimed materials requires a design method for creatively reassembling distinct building components. Generative artificial intelligence (AI), capable of creating content based on learned data patterns, offers a new possibility for conceptual design in architecture. The study explores generative AI-aided design strategies for creatively (re-)assembling reused building materials and proposes an innovative circular design framework. We investigate the role of generative AI for exploring image-to-image and text-to-image generators in creatively managing and imaginatively applying building component databases for circular design. We show the potential of generative AI for supporting human creativity in the proposed co-design approach to forge a new aesthetic paradigm in architecture focused on the creative amalgamation of distinct elements.
  • Empirical Quantification and Prediction of Building Component Lifespans with Graph Neural Networks
    Item type: Conference Paper
    Wu, Pei-Yu; Mundt-Petersen, S. Olof (2025)
    Extending the service life of existing buildings and their components is essential for slowing resource loops in circular construction. While overall building lifespans are relatively well understood, the actual service lives of individual components remain underexplored due to limited empirical data. This study addresses this gap by analyzing and modeling building component lifespans using real-world damage investigation records and condition assessments from Swedish buildings. Descriptive statistics reveal a wide range of lifespans across six major component categories, highlighting materials linked to damage from poorly designed construction details as a critical factor. Additionally, most residential components have markedly shorter lifespans-about 25–30 years-compared to non-residential ones. To predict component service lifespans, we configure building envelope deterioration factors as a generic acyclic graph and employ Graph Neural Networks for node-level regression. Among the tested models, Graph Convolutional Network achieved the highest performance (R² = 0.83), outperforming GraphSAGE and Graph Attention Network due to its effective uniform aggregation across neighboring nodes. Our findings provide a high-granular, data-driven approach to estimating component lifespans, which can enhance predictive maintenance strategies and optimize the reuse potential of reclaimed materials in their next lifecycles.
  • Predictive modeling and estimation of moisture damages in Swedish buildings: A machine learning approach
    Item type: Journal Article
    Wu, Pei-Yu; Johansson, Tim; Mundt-Petersen, S. Olof; et al. (2025)
    Sustainable Cities and Society
    Identifying potential moisture damage is crucial for maintenance practices and assurance of well-being of occupants. However, due to limited information availability and standardization, assessing damage prevalence on the building stock scale remains understudied. By combining investigation records and building databases, this study leverages data analytic techniques and machine learning modeling to characterize damage pathology and predict its occurrence in Swedish buildings. The interrelationships between damage-specific attributes and their associations with building parameters of several damage types were analyzed using feature selection, forming the basis for developing predictive models. Results show that multilabel classifiers outperform binary classifiers for every damage type, with lead tree ensemble models achieving minimum average AUCPR and F2 of 0.85 for microbial growth, 0.87 for deformation, 0.91 for odor, and 0.95 for water leakage. The identified patterns were interpreted and verified against descriptive statistics. The binary relevance models estimate that one-third of school buildings, 20 % of commercial and office buildings, and 15 % of residential dwellings in regional building stock contain moisture damage. These findings advance the quantification of moisture damage by providing new knowledge and approaches for appraising moisture damage likelihood at aggregated and individual building levels, thereby aiding in moisture safety evaluations and preventive maintenance efforts.
  • Probabilistic Distributions of Moisture Damages in Swedish Buildings
    Item type: Conference Paper
    Wu, Pei-Yu; Johansson, Tim; Mundt-Petersen, S. Olof; et al. (2025)
    Lecture Notes in Civil Engineering ~ Multiphysics and Multiscale Building Physics
    Moisture damages lead to significant costs and impact indoor environments negatively. Identifying their occurrence patterns is crucial for the implementation of preventive or mitigative measures, though it remains highly challenging due to the complex interplay of multidimensional variables. This study aims to identify primary moisture damage profiles in Swedish buildings and evaluate their prevalence. Using data-driven analytical and visualization techniques, 2,100 complex moisture-related damage records between 2014 and 2020 with information on building parameters and damage specifics from Sweden were examined in multivariate analysis. The association analysis reveals varied relationships among factors related to damage, with damage types distributed proportionally across damaged components, causing components, sources, building phase, and responsible actors. The presence patterns of damage differ significantly by building types, yet they are generally reported more frequently in buildings built during the 1960-1980 and 2000-2020. Prevalent moisture damages in Swedish buildings include microbial growth and deformation at the building envelopes and roof, odor in indoor environments caused by wind-driven rain, and indoor or outdoor humidity. Additionally, these damages appear more often in buildings constructed with non-ventilated crawlspaces, wood, concrete, brick structures, facade, exterior walls, and non-ventilated cold attics. The characterized moisture damage patterns and estimated frequency enhance the understanding of their occurrence and associating factors.
  • A unified intelligence-augmented framework for building audits via physical and virtual inspection
    Item type: Journal Article
    Wu, Pei-Yu; El-Assady, Mennatallah; De Wolf, Catherine (2026)
    Expert Systems with Applications
    Building audits are essential for informed decision-making in maintenance, renovation, and end-of-life planning. However, current practices remain predominantly manual and time-consuming due to fragmented data, limiting resource-efficient management of existing building stock. This paper presents a unified, intelligence-augmented framework designed to enhance the efficiency and reliability of both physical and virtual building inspection workflows. Five core design principles of adaptability, accessibility, affordability, acceleration, and alignment are derived from a multi-phase formal analysis to guide the development of the R2PIVS pipeline, which transforms the existing audit process into six modules: retrieval, reality capture, prediction, interaction, visualization, and summarization. The framework leverages human-AI collaboration in key building inventory tasks, including geometry measurement, visual assessment, and hazard estimation, through interactive annotation, model refinement, and output validation. Findings from expert elicitation studies indicate that the proposed application schema is promising for improving the efficiency and scalability of existing workflows. By aligning machine learning capabilities with domain-specific requirements, this research lays the foundation for a human-in-the-loop building audit system that enables standardized inspection and inventory information management to support circular construction practices throughout the building life cycle.
Publications 1 - 5 of 5