Aerodynamic Modelling and Rejection of Wind Disturbances for Multirotor Unmanned Aerial Vehicles

Abstract

Multirotor unmanned aerial vehicles, or UAVs, are being introduced into the New Zealand agricultural and forestry industries to increase efficiency and safety whilst reducing costs. UAVs are being developed to perform complex, interactive tasks within their environment which requires a high level of station keeping (closely maintaining a desired position). Wind gusts have been identified as a limiting factor to current UAV operations as they degrade stability, station keeping and flight times. The purpose of this research is to investigate the aerodynamics of UAVs with the intention of identifying areas in which wind disturbance rejection can be improved through simple, passive methods, without significantly increasing UAV mass. Aerodynamic force and moment models with inputs of wind speed, propeller angular speed and angle of attack were developed from static experimental wind tunnel testing of an octocopter and its individual isolated propellers to gain valuable insight of a UAV’s aerodynamic behaviour in wind. Simple, generalized aerodynamic coefficient functions were found to effectively model forces and moments for a full UAV as well as isolated propellers. A novel summed UAV rotor and body aerodynamic model was developed to investigate modelling efficiency. It was compared to the full aerodynamic model and general trends matched extremely well. However, significant unaccountable and changing thrust losses were discovered when propellers are in a multirotor configuration. A discrete gusting system was developed for a wind tunnel to enable the repeatable, free-flight testing of a UAV when subjected to wind disturbances representative of New Zealand conditions. Using the aerodynamic models developed, body drag was identified as a substantial disturbance contributor warranting further investigation. A drag reducing shroud was developed and compared to the baseline octocopter in a series of free-flight experiments, utilizing the discrete gusting system. A series of trials saw an average 13% decrease to maximum station keeping error, proving simple, passive methods of wind disturbance rejection are possible.