Abstract This article presents a numerical approach to predict the pyrolysis and heat transfer in spruce wood during fire resistance tests using Computational Fluid Dynamics (CFD) in ANSYS Fluent. Fire safety constructions often involve complex three-dimensional geometries, which pose significant challenges for numerical modeling in terms of computational effort and stability. To address this, the presented material model is intentionally kept as simple as possible, prioritizing practical usability and robustness for engineering applications such as fire resistance tests for fire protection doors, windows, and fire dampers. The model simulates the key thermal processes relevant to fire exposure, including the dehydration of moist wood, its transition to char, and heat transfer through the wood bulk. While it does not attempt to capture detailed chemical kinetics, it reliably predicts temperature evolution, making it a valuable tool for the development of fire-resistant components. This simplification also reflects the originality of the study, which focuses on practical usability and applicability in real engineering scenarios. Key temperature-dependent properties of wood and char were experimentally determined through thermal diffusivity Laser Flash Analyzer (LFA), specific heat capacity Differential Scanning Calorimetry(DSC), thermal expansion Dilatometry(DIL), and mass loss Thermogravimetric Analysis (TGA). These properties were used as input data for numerical simulations. In the simulation, wood is modeled as a porous medium, solving heat and mass transfer equations to predict temperature, pressure, and gas composition during pyrolysis. The Arrhenius Law model describes the chemical reactions, capturing wood degradation into volatile gases and char. The model was validated by small-scale warm-up tests up to 400°C, demonstrating its ability to predict internal temperature evolution. The Lee model was incorporated to account for steam re-condensation, which significantly influences thermal behavior. Validation against real-scale fire resistance tests followed standard EN 1363-1. The model accurately predicted temperature variations at different depths within a large spruce wood sample. While effective for predicting fire resistance, deviations arose due to the exclusion of heat release from burning wood and the lack of modeling for the destruction of the char layer at high temperatures. This model provides a cost-effective tool for evaluating fire-resistant wooden structures, particularly in partially wooden elements like timber stud walls with gypsum cladding, where ignition and structural failure do not occur.
Kitzmüller et al. (Mon,) studied this question.