Numerical study of ice particle shedding: Interpolation methods and 2D trajectories

Flying in icing conditions can be dangerous and de-icing systems can raise safety concerns: ice blocks shed can impact elements like the engines. Numerical prediction of these blocks trajectories still lack of data, mainly due to the random aspect of ice shapes and initial positions. The first step will be to determine the flow field around an aircraft geometry using the RANS approach and the Finite Volume Method with the open source SU2 CFD code. Using an interpolation method, we will next be able to compute the velocity at every spatial location inside the discretized computation domain. The aerodynamic behaviour of typical spherical and flat plate shaped ice blocks will be introduced into that velocity field to follow their trajectories along the airplane. In this paper, we will focus on interpolation and compare the effect of interpolation methods available on MatLab on the computation, in first attempt, of 2D trajectories. The aspects we will focus on will be the trajectory paths and the computational resources -memory and execution time- used for each interpolation method. The linear and natural methods are faster than the nearest one and compute similar trajectory paths, but with some differences. Future work will be to implement it in 3D and the use of a Monte-Carlo method to create the probability map will be the last step. It will lead us to find some critical zones with high chances of ice transit and thus help designers to find a safe location to position the aft-mounted engines among others.