Robust and Efficient Prediction of the Collection Efficiency in Icing Accretion Simulation for 3D Complex Geometries Using the Lagrangian Approach I: an Adaptive Interpolation Method Based on the Restricted Radial Basis Functions

Liang Xie,Peizhe Li,Hang Chen,Hong Liu
DOI: https://doi.org/10.1016/j.ijheatmasstransfer.2019.119290
IF: 5.2
2020-01-01
International Journal of Heat and Mass Transfer
Abstract:As a preliminary step in the icing accretion simulation, prediction of the collection efficiency is important for the icing analysis. At the present time, there are two different ways to accomplish this task, namely, the Eulerian and Lagrangian approach. The former could be easily implemented to 3D complex geometries and widely used right now because it can be obtained together with the Eulerian computation of the flow field. But it requires very much computational costs include both memory and CPU time. On the other hand, the latter has great potential to achieve high efficiency. However, its efficiency has been demonstrated only in simple cases. To facilitate the implementation of the Lagrangian algorithm for 3D complicated configurations, some new techniques are proposed or introduced into the Lagrangian framework in the current research. The first is a new δ form of the local collection efficiency. In this new form, four water droplets are placed very close together to ensure that all of them could hit the wall, so that the collection efficiency could be evaluated robustly for 3D cases. The second is that the adaptive mesh refinement (AMR) algorithm is introduced into the computation of droplets trajectories. The AMR method could capture the droplets impingement limits correctly in the Lagrangian approach. The third is that the interpolation algorithm based on the radial basis functions (RBF) is employed in the present paper to transfer the collection efficiency evaluated at the impingement positions to the CFD (computational fluid dynamics) mesh. To prevent the un-impinged area from being immersed by the interpolation, the restricted RBF method developed by the authors for the mesh motion is also adopted here. The above simple but useful techniques are synthetically used in the Lagrangian method and lead to an algorithm which is suitable for 3D complex configurations. Some cases include several 3D complex models are used here to validate the performances of the proposed approach. These examples show that the algorithms developed in our research can be efficiently applied to 3D complicated shapes and provide reliable results.
What problem does this paper attempt to address?