REV-based analysis of non-uniform heat transfer in irregular geometries of precast concrete sandwich wall panels
Year: 2026
Type: article
Abstract: The thermal performance of buildings is largely determined by the weakest point in the envelope. Precast concrete sandwich wall panels are a new type of high-performance, multi-functional integrated component that has been widely adopted. However, local heat loss caused by metal connectors represents a significant performance limitation for such panels. For irregular geometries such as truss-type connectors, conventional analysis methods are not suitable. A methodology that simultaneously balances cost, accuracy, and the ability to capture localized heat transfer characteristics is currently lacking. To address these challenges, this study proposes a representative element volume (REV)-based heat transfer analysis method. Based on the heat transfer characteristics of the connector joints, three distinct heat transfer zones were defined, clarifying both the influence scope of the thermal bridge and the methodology for quantifying heat transfer. Through a three-level orthogonal analysis of ten design factors, the influence relationships between significant factors and critical parameters in the REV method were established. Consequently, simplified computational models were developed for four critical parameters. The models demonstrated robust performance, with R² values exceeding 0.9. The mean relative errors between model predictions and experimental results (3.83% for this study and 2.45% for comparable studies) demonstrate the method’s accuracy over existing standard approaches such as the linear heat transfer coefficient method and full-scale modeling, while eliminating the need for resource-intensive numerical simulations. This work provides a foundational theoretical framework and an advanced thermal modeling methodology for optimizing precast concrete sandwich wall panels.
Language: English
Subfield: Building and Construction
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $4,220
Award: G4797420149
OpenAlex ID: W7214318549
