中文摘要：

为了研究LNG泄漏扩散过程及危害,建立了引入时间参数的高斯烟羽混合模型,利用MATLAB工具对LNG泄漏扩散过程进行动态模拟,解决了高斯烟羽模型不能模拟连续泄漏源泄漏初期浓度分布的问题。提出了非点源高斯烟羽混合模型,可预测液池、大孔等非点源的泄漏扩散过程,并利用Burro 9号LNG泄漏扩散试验进行模型验证。研究了风速、大气稳定度等对LNG泄漏扩散所形成的危险区域的影响,结果表明：风速对LNG泄漏扩散的影响显著,风速越大,扩散越快,扩散达稳定后所形成的危险区域面积越小;大气越稳定,扩散越慢,危险区域面积越大。

英文摘要：

This paper is aimed at simulating the process of LNG leak- age and dispersion as well as analyzing the impacts of the key param- eters such as wind velocity on the hazardous area. For the above pur- pose, we have established a time-integrated mixed Gaussian Plume Model for simulating the dynamic LNG dispersion process via a soft- ware MATLAB. In the aforementioned model of ours, the continuous leakage source should be regarded as a limited number of transient leakage sources with time parameter being introduced based on the assumption that one transient leakage source is likely to be produced in one second. The total concentration of natural gas in the field is the sum of the concentration produced by every transient leakage source. Therefore, this model we have developed is endowed with a striking advantage of simulating the initial stages of dispersion and de- scribing the process of LNG dispersion intuitively while Gaussian Plume Model could only simulate the concentration distribution of steady state. Thus, the simulation results of ours demonstrate that LNG concentration tends to increase gradually and then keeps steady as the time goes on. Besides, the concentration of natural gas at the point near the source may reach a steady state more quickly than those far away from the central area. Thus, the proposed non-point source Mixed Gaussian Plume Model can help to simulate the disper- sion process of the non-point source, such as pool and big hole, which should be regarded as lots of point sources uniformly distributed in the leakage area, and the total concentration of natural gas in the field should be the sum of the concentration of natural gas produced by every source point. The simulated non-point source model we have proposed has been validated actually with the experimental data from the Burro 9 experiment. It indicates that non-point source Mixed Gaussian Plume Model is superior to the point source Mixed Gaussian Plume Model, in which the impacts of the wind velocity and the at- mospheric stability on