Study of the trajectories of ice shed by deicing system around aircraft engine

Abstract

The problem of ice accretion caused accidents and incidents to aircraft over the past decades. Solving the problem of ice accretion employing de-icing and anti-icing devices will remove the accumulated ice but the problem of unknown trajectories of the detached particles appears. The flown particles represent a great hazard on the aircraft which yields risk depending on the vulnerability of aircraft parts. The vulnerable aircraft parts are the wing, the rear fuselage, the stabilizers, and the rear-mounted engines. In order to mitigate the risk, a study of the trajectories of those particles is introduced. The objective of this research is to study the trajectories around a wing by changing the angle of attack and the sweepback angle. The goal is to calculate the minimal number of ice trajectories to correctly predict a footprint map at the inlet section of the engine using the Monte-Carlo method. A numerical approach is used to accomplish this study. The random trajectories of the ice particles are calculated using a 3D Panel Method (3DPM) flow field around the wing. To determine the zones behind the wing where the ice particles have the most passage probability, a Monte-Carlo method is utilized. In this research, the calculations are done through a probabilistic study of the footprints to determine their probability distribution's shape. Once the shape is known, a normality test is done on the shape of the Probability Distribution Function (PDF) called the Kolmogorov-Smirnov test. After determining the shape and the type of the PDFs, a study on the mean and variance for every PDF is done to check the minimal number of trajectories to fulfill the Monte-Carlo method. The 3DPM flow field is validated against the literature as well as the footprint distribution behind the wing. The effect of the angle of attack, as well as the sweepback angle on ice particle trajectories, is shown. The increase of the angle of attack shifts the trajectories upward while the sweepback makes the footprint map less noisy. Finally, 500 trajectories were found enough to predict a footprint map.

Date	12 Dec 2019
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	François Morency (Supervisor)