Computational Fluid Dynamics (CFD) is used to predict the dispersion of a gasoline vapour cloud. The model was validated using the results of the analysis of Buncefield incident (2005) in the UK and a new method of simulation that involves the interaction of liquid fuel and air during the dispersion was introduced. The Caribbean Petroleum Corporation (CAPECO) incident (2009) was simulated using the developed CFD modelling method.

The results of CFD analysis of this incident focused on the release of flammable liquid fuels that was followed by the formation of vapour clouds. The effect of wind speed and the gasoline evaporation surface area were analysed.

The CFD simulations showed that the main factors affecting the dispersion and behaviour of the vapour clouds are release surface area, wind speed, topography, turbulence and obstacles in the area. The simulation results showed similar behaviour to that observed in the real-life incidents in terms of the vapour cloud behaviour and the extent of damage.

This work used the realizable $\text{k}-\epsilon$ model which was previously validated for the dispersion of dense gases. This turbulence model was helpful in the study of the wind speed (nil to low) characteristics of the CAPECO dispersion accident and the dispersion of the liquid gasoline droplets was successfully simulated using ANSYS FLUENT's built-in Discrete Phase Model (DPM). This work simulated evaporation of gasoline liquid and droplet dispersion in a wide area and from a large source with the presence of obstacles. Dispersion of the gasoline droplets was successfully accounted for by the inclusion of the DPM sub-models for cloud dispersion, namely, the stochastic tracking and the Discrete Random Walk model (DRW). It could be concluded that this work validates the feasibility of using the DPM sub-model application for the simulation of volatile liquids evaporation and dispersion through a domain that includes several obstacles and high levels of congestion.

计算流体动力学（CFD）用于预测汽油蒸气云的扩散。该模型使用英国Buncefield事件（2005）的分析结果进行了验证，并引入了一种新的模拟方法，该方法涉及分散过程中液体燃料和空气之间的相互作用。加勒比海石油公司（CAPECO）事件（2009）是使用已开发的CFD建模方法进行模拟的。

CFD分析此事件的结果集中于释放易燃液体燃料，随后形成蒸气云。分析了风速和汽油蒸发表面积的影响。

CFD仿真显示，影响蒸气云散布和行为的主要因素是释放表面积，风速，地形，湍流和该区域的障碍物。仿真结果表明，就蒸气云行为和破坏程度而言，其行为与真实事件中观察到的行为相似。

这项工作使用了可实现的 $\text{ķ}-\epsilon$先前已验证过的用于浓密气体扩散的模型。该湍流模型有助于研究CAPECO弥散事故的风速（从零到低）特征，并且使用ANSYS FLUENT的内置离散相模型（DPM）成功地模拟了汽油液滴的弥散。这项工作模拟了汽油液体的蒸发和液滴在大范围内的扩散，并从有障碍物的较大来源进行了模拟。通过包含DPM子模型以进行云扩散，成功地解决了汽油液滴的扩散问题，即随机跟踪和离散随机游走模型（DRW）。