The Computational Cost and Accuracy of Spray Droplet Collision Models 2019-01-0279
This study focuses on Lagrangian spray models that are commonly used in engine CFD simulations. In modeling sprays, droplet collision is one of the physics that must be accounted for. There are two main parts of droplet collision models for sprays -- detecting colliding pairs of droplets and predicting the outcomes of these collisions. For the first part, the focus is on the efficiency of algorithm. We present an implementation of the arbitrary adaptive collision mesh model of Hou and Schmidt (IJMF 2006), and examine its efficiency in dealing with large simulation problems. Through theoretical analysis and numerical tests, it is shown that the model exhibits pseudo-linear scaling of cost with respect to the number of parcels in the sprays. Regarding the second part, we examine the variations in existing phenomenological models used for predicting binary droplet collision outcomes. A quantitative accuracy metric is used to evaluate the models with respect to the experimental data set. To provide a holistic perspective, three different experimental datasets which contain over 2000 data points are collated, and used in the evaluation. It is found that the accuracy of the collision outcome prediction is between 65% to 70%, which shows room for future improvement.
Arpit Agarwal, Yue Wang, Long Liang, Chitralkumar Naik, Ellen Meeks