气液两相流中液滴识别及体积计算

Identification and Volume Calculation of Liquid Droplets in Gas-Liquid Two-Phase Flow

气液两相流形态受到两相性质、气流和壁面等条件的影响,通常液滴在空间上的分布结构丰富多变,使得在空间中识别液滴及计算液滴体积变得十分复杂。为解决这些问题,将带有噪声的基于密度的空间聚类算法(Density-Based Spatial Clustering of Applications with Noise,DBSCAN)和光滑粒子流体动力学方法(Smooth Particle Hydrodynamics Method,SPH)结合使用。该结合方法可利用DBSCAN算法来解决液滴识别问题,并使用SPH方法计算液滴体积。经过对比验证,计算结果表明,该方法可以正确识别气液两相流中液滴数量及其体积计算,计算结果最大误差不超过真实值的2.2%,相较于目前已有的体素法更优,并且适用于液滴大形变情况。通过对横向射流雾化过程计算分析,表明该结合方法可以正确反应液体破碎的雾化程度及雾化效果。

Gas-liquid two-phase flow morphology is affected by the properties of the two phases,airflow,and wall conditions.Usually,the distribution structure of droplets in space is rich and varied,making it extremely complex to identify and calculate the volume of droplets in space.To solve these problems,we combine a density-based spatial clustering algorithm with noise(DBSCAN)with a smooth particle hydrodynamic method(SPH)that contains noise.This combined method can use the DBSCAN algorithm to address the problem of droplet identification and use the SPH method to calculate the volume of droplets.After comparative verification,the calculation results show that this method can correctly identify the number and volume of droplets in gas-liquid two-phase flow,and the maximum error of the calculation results is no more than 2.2% of the true value,which is better than the existing voxel method and suitable for large deformation of droplets.Through computational analysis of transverse jet atomization processes,this combined method can accurately reflect the degree of liquid fragmentation and atomization effect.