Physics-based model for wood charring under the standard temperature–time curve

Charring is a critical aspect of the burning of wood, as it influences the fire dynamics, and leads to a weakening of timber which compromises the structural integrity of buildings. This study presents a one-dimensional physics-based model for wood charring with extensive comparison to experimental data and other models. The chemical and thermophysical parameters are taken from the literature, while the boundary conditions are calibrated on two ISO-834 furnace experiments. The model is then validated against 67 experiments from 6 different studies in the literature to corroborate its validity well beyond the calibration cases. Temperature profiles at different depths yield an average error of 12.5% against thermocouple data, and correctly predicts the surface regression of the sample. The progression of the char depth yields an average error of 8%. The model is compared against the linear correlation of the Eurocode, showing that the latter underpredicts the char depth with an error of 19%. In this paper, the physics-based model is used to analyse the dependency of the charring rate on the heating rate, and to show that charring is not instantaneous but occurs over a range of 200 ° C . The effect of moisture content and density on the charring rate are also investigated. At higher moisture contents the charring rate is lower, due to the heat sink during evaporation, slowing down pyrolysis. At lower densities the charring rate is higher. It is shown that a well-characterised physics-based model, built on established parameters from the literature is valid for a number of different experiments. It also captures the transient behaviour of charring and the influence of critical parameters, like moisture content. This study compares the performance of the physics-based model to the one of a neural network model, a statistical model and the Eurocode. The physics-based model proves to best predict the behaviour of timber charring, compared to the other models, especially its dependency on moisture content.