S Bradley
lvl.1
United States
Offline
|
American Institute of Aeronautics and Astronautics 21 VI. Conclusions The paper presents a comprehensive method for the design of quiet propellers, which is based on a Multidisciplinary Design Optimization (MDO) approach. The design of a quiet propeller includes usually a high number of design variables (69-70 design variables in the above described example) and constraints. Thus in order to cope with the complexity of this design problem, three different optimization schemes are used that enable the design process to reach a global optimum. The designer is responsible for the definition of the cost function and constraints. In addition, the designer decides on transferring from one optimization scheme to the other. Often, the design trends of a quiet propeller contradict the design trends of an efficient propeller. Thus a trade off is required in order to arrive at a practical design. The efficiency of the propulsion system does not depend solely on the propeller, but it is affected significantly by the motor efficiency and the vehicle characteristics. Thus it is important to include the motor characteristics in the design process. Based on the example of designing a quiet propeller for an electric Mini-UAV, the following conclusions can be drawn: a. For a given number of blades, the main mechanism of reducing the propeller noise involves a reduction of the propeller’s rotational speed. b. Defining the propeller SPLA as the cost function, without applying additional constraints, may result in an impractical design of blades having a very large chord and high required power. In the present example the loitering time for such a design was reduced from 160[min.] in the case of a maximum loitering time design (without considering any other constraints), to only 25[min.] for a very quiet propeller deign (without apply any performance constraints). c. Imposing power constraints, namely imposing a limit of the extracted power from the battery, increases the propeller noise significantly (compared to a design without such constraints). In the case of a two-bladed propeller, limiting the power to 75[W] results in an increase of 10.3[dBA], while a stricter power limit of 65[W] results in an increase of 12.3[dBA]. This increase in noise is mainly due to the increase in the rotational speed that leads to an improvement in the efficiency of the propulsion system. d. Imposing stress constraints leads to an increase in the cross-sectional thickness and the rotational speed. Due to the increase in the rotational speed the SPLA is increased as well (by 1~3[dBA] for a two-bladed propeller). It has been shown that by introducing a cone angle, the aerodynamic bending moment is reduced and the optimal propeller’s SPLA is decreased by 0.1~0.3[dBA]. e. As the number of blades increases, the noise signature is decreased. The reduction in noise is a function of the power constraint and can reach 10[dBA]. f. The influence of stress constraints on the optimal design depends strongly on the number of blades. g. During a flyover the noise of a UAV varies as a function of the location of the vehicle relative to the observer. The maximum SPLA is obtained for a receding vehicle, shortly after it passes the zenith point. Nevertheless, optimization for a minimum noise at the zenith point reduces the noise signature for most of the flyover case, especially along the noisier parts of the path. h. For most cases, as the performance constraint becomes more strict (smaller power is allowed), the noise signature increases. Only for an approaching vehicle, at certain distances from the observer, this trend can be reversed. i. The far field loading noise presents the main contribution to the propeller noise. The contributions of the thickness noise and near field loading noise are usually much smaller. The results of the present paper emphasize the importance of a multidisciplinary approach during the design of a quiet propeller. It is important to address all the disciplines simultaneously along the design process. Ignoring one of these disciplines may lead to an impractical design. |
|