基于统计学特征的密集型气液雾化随机模型优化方法研究

Improvement of Stochastic Model Based on Statistical Characteristics for Dense Air Atomization

关键词:  空气雾化流场  数值模拟  随机浸入体模型  大涡模拟方法  亚网格动能粒子追踪法

Key words:  Air atomized flow field  Numerical simulation  Random immersed model  Large eddy simulation method  Sub-grid kinetic energy particle tracing method

摘要:

在高气液动量比的空气雾化流场中，液滴和液丝从液核剥离的过程具有高自由度、分布密集的特点，传统理论模型难以对其准确预测。本文对结合大涡模拟方法的随机雾化模型进行优化，在初始雾化过程，提出使用液滴统计平均温度来表征液滴碰撞统计学特性的改进方法。液滴的统计平均温度分别采用气液相对动能模型的粒子追踪法和亚网格动能模型的粒子追踪法。研究表明，使用改进的气液雾化随机模型预测密集型空气雾化流场，大幅改善了传统雾化随机模型在初始雾化区域过预测的缺陷，平均动能的相对误差为15.5%，平均索特尔直径的相对误差为7.2%，与未改进前的模拟结果相比，误差降低了41.1%和15.0%。此外，本文还探究了喷雾张角模型对雾化流场预测结果的影响，分别将实验所得经验公式法、相界面气液动量平衡所得模拟法与亚网格动能模型的粒子追踪法结合。结果表明喷雾张角经验公式预测结果更为准确，在平均索特尔直径预测方面准确性提高了17.3%。

Abstract:

In the air atomization flow field with high gas-liquid momentum ratio， the stripping of droplets and filaments from the liquid core has the characteristics of high degree of freedom and dense distribution. Therefore， the traditional model is difficult to predict accurately. The traditional stochastic atomization model combined with large eddy simulation is improved. In the primary atomization process， an improved method is proposed to characterize the statistical characteristics of droplet collision by using statistical mean temperature. The statistical mean temperature of droplet is calculated by particle tracking method of gas-liquid relative kinetic energy model and sub-grid kinetic energy model respectively. The results show that the improved stochastic model greatly improves the over-prediction of the traditional stochastic model in the primary atomization region. The mean kinetic energy relative error is 15.5% and the mean sauter diameter relative error is 7.2%， which is 41.1% and 15.0% lower than the simulation results before improvement. In addition， the influence of the spray angle model on the prediction of the atomization field is also explored. The empirical expression method and the simulation method which is derived by gas-liquid momentum balance at the interface are combined with sub-grid kinetic energy model. It shows that the improved stochastic model using sub-grid kinetic energy model and empirical spray angle expression is more accurate. In terms of prediction of mean sauter diameter， the accuracy is improved by 17.3%.