Focusing on convective heat transfer modelling upward flame spread scenarios, different methods for calculating the convective heat fluxes are explored. These approaches involve: 1) calculating convective heat transfer based on the stagnant film theory; 2) using a simplified approach based on the stagnant film theory; 3) using Nusselt correlations for natural convection; and 4) modelling convection based on the effective (i.e., laminar plus sgs) thermal diffusivity. Large eddy simulations (LES) of a series of Single Burning Item (SBI) experiments with Medium Density Fiberboard (MDF) and plywood walls are performed. Even though discrepancies are observed, the approach which considered the sgs effects performs overall better and is able to predict both the initial flame growth period and the decay phase with reasonable accuracy (i.e., deviations in the peak HRR of approximately 5$\text{\%}$ and 34$\text{\%}$ for the MDF and plywood tests cases, respectively). The predicted total heat fluxes are, in most cases, in fairly good agreement with the experiments (i.e., deviations of up to 15$\text{\%}$ and 20$\text{\%}$ for the MDF and plywood tests cases, respectively). The predictions with the other methods are far less satisfactory. Overall, it is demonstrated that accurate convective heat transfer modelling can be important in flame spread scenarios.

着重于对流传热建模向上的火焰扩散场景，探索了计算对流热通量的不同方法。这些方法包括：1）基于停滞膜理论计算对流换热；2）使用基于停滞薄膜理论的简化方法；3）使用Nusselt相关性进行自然对流；和4）根据有效（即层流+ sgs）热扩散率对流建模。使用中密度纤维板（MDF）和胶合板墙进行了一系列单一燃烧项目（SBI）实验的大型涡流模拟（LES）。即使观察到差异，但考虑sgs效应的方法总体上效果更好，并且能够以合理的精度预测初始火焰生长周期和衰减阶段（即，$\text{％}$ 和34$\text{％}$分别用于MDF和胶合板测试用例）。在大多数情况下，预测的总热通量与实验相当吻合（即，偏差最大为15$\text{％}$ 和20$\text{％}$分别用于MDF和胶合板测试用例）。使用其他方法的预测远不能令人满意。总体而言，这证明了精确的对流传热模型在火焰扩散场景中可能很重要。