Numerical and experimental investigation of ice shedding
The shedding of ice accreted on aircraft surfaces poses a serious threat to flight safety as it can cause severe damage to downstream aircraft components such as aft-mounted engines. While there is strong need for ice shedding simulation tools to support aircraft icing analysis and certification, currently available ice shedding analysis methods may not predict ice fragment trajectory accurately due to the inability to simulate the random nature of ice shedding events and the lack of experimental aerodynamic data for ice fragments. In addition, most of the simulation tools have not been validated with experimental trajectory data. Both experimental and numerical investigations of shed ice trajectory were conducted as part of the continuing development of an ice trajectory simulation tool at Wichita State University. The research discussed in this thesis involved wind tunnel experiments conducted to obtain experimental aerodynamic coefficients for simulated ice fragments, numerical ice trajectory analyses, Monte Carlo simulations to evaluate the risk of ice fragment impact on critical aircraft components, and trajectory experiments performed to validate the ice trajectory simulation tool. Wind tunnel experiments were performed in the WSU 7-ft x 10-ft wind tunnel facility to obtain an aerodynamic database of force and moment coefficients for a 12-inch square flat plate with 0.4-inch thickness, a 10-inch diameter disk with 1-inch thickness and a 18-inch span glaze ice fragment with symmetric horns computed with the LEWICE ice accretion code. The experiments were conducted at airspeed of 160 mph with varying Euler angles (yaw, pitch and roll). The aerodynamic coefficients of the three fragments tested demonstrated considerable sensitivity to fragment shape and orientation. The experimental data were incorporated into the WSU trajectory analysis code developed. The ice shedding analyses were conducted with two simulated ice fragments for a business jet aircraft using the WSU trajectory code. The two simulated ice fragments include a disk and a glaze ice fragment with symmetric horns released from the aircraft nose and from the antenna on the top of the fuselage, respectively. Monte Carlo simulations were performed to compute probability maps of trajectory footprints at the engine inlet plane. The analysis demonstrated the effect of the shed location, initial fragment orientation and aircraft angle of attack on ice fragment trajectories. The results for Monte Carlo simulations performed with the disk fragment showed that the fragment would have less than a 0.15% chance of impacting the engine. For the glaze ice fragment, it was found that the probability of a collision between the fragment and the engine was as high as 24.2%, depending on the aircraft’s pitch and yaw angles. Next, an experimental methodology was developed to obtain the trajectories of simulated ice fragments released in a tunnel airstream for the validation of the ice fragment trajectory code. The tested fragments were a 6-inch square flat plate with 0.4-inch thickness, a rectangular flat plate measuring 12- inch long by 6-inch wide by 0.4-inch thick, a 12-inch span single horn glaze ice fragment and a 12-inch span double horn symmetric glaze ice fragment. Experiments were performed at airspeed of 160 mph in the WSU 7-ft x 10-ft wind tunnel facility. High-speed video cameras were employed to record the fragment’s trajectories at 500 to 1,000 frames per second. The coordinates of a fragment’s trajectories were determined from high speed images with the help of gridded vinyl sheets attached to one of the wind tunnel side-walls and to the ceiling. The fragment trajectories showed considerable sensitivity to the ice fragment shapes and their initial pitch angles at the moment of release. The flat plates with initial pitch angle of 0° experienced considerable rotation as they moved downstream, while the flat plate with initial pitch angle of 90° traveled downstream in a nearly straight path without rotation. Two cases of the single horn glaze ice fragment were sensitive to initial orientation and exhibited oscillatory rotations with respect to their spanwise axis. Two cases of the double horn symmetric glaze ice fragment resulted in similar trajectories, however experienced considerably different speed. Finally, the experimental trajectories were compared with analytical trajectories computed with the WSU trajectory code. Good agreement was demonstrated between the experimental and computed trajectories in most cases. However, the analytical results for the flat plate cases with significant rotations were more than 10% different compared to the experimental data due to the effect of the plate large rotation speed on the aerodynamic forces and moments acting on the plates.
Thesis (M.S.)--Wichita State University, College of Engineering, Dept. of Aerospace Engineering.