CFD Effect of Propellers Interference on the Thrust Variation in the Proximity of Fixed and Horizontal Moving Obstacle

Abstract

Quadcopters have gained widespread popularity in various sectors including military and civilian applications. They are used for tasks such as cargo transportation, military surveillance, and public aerial photography. However, when flying near obstacles or moving objects, quadcopters may encounter external forces that can cause flight instability. This research involves the use of Computational Fluid Dynamics (CFD) simulations to study the flow dynamics interference of two Phantom 3 drone propellers as they hover near fixed and moving obstacles. The simulations use a tetrahedral mesh with multiple reference frames (MRFs) and are validated by comparing the results to experimental and theoretical data. The simulations are conducted at different altitudes (0.1 m, 0.2 m, 0.5 m, and 1 m), rotational speeds (3000 rpm, 5000 rpm, and 9550 rpm), and with the obstacles moving at different velocities (5 m/s, and 10 m/s), in order to understand the aerodynamic effect of these variables on the performance of the propellers. When hovering (9550 rpm) at 0.1 m above a fixed obstacle, the results have shown first an increase in the lift by 9.28% for an isolated propeller, and then an increase by just 3.94%, and 4.39%, for the first and the second propeller, respectively, when placed next to each other (29 cm spacing). The variation in thrust is also reduced by half when the propellers are moving far from the wall (at 0.2 m height) compared to a sole propeller thrust. Finally, the moving wall below the propellers at several velocities proved to also have a significant effect on the thrust variation of the propellers at low hovering speeds.

A comprehensive study on the aerodynamic influence of stationary and moving obstacles on an isolated phantom DJI 3 UAV propeller

“…To build upon and bolster our earlier studies [27,37,38], mesh independence and convergence studies were performed for this computational model using the methodology presented in Figure 4a. The tests included five variants, as they allowed one to capture the trend and draw conclusions.…”