扑旋翼飞行器机构动力学建模与功率优化

The classical gear-linkage mechanism has been adapted in the state-of-the-art flapping wing rotor (FWR) micro aerial vehicles (MAV). The driving motor needs to provide power against both inertia force and aerodynamic forces of the flapping wings. This leads to a low efficient mechanical system. The present study is to evaluate the efficiency improvement by adding a spring in the mechanism through theoretical, numerical simulation, and experimental methods. Firstly, the dynamic model for an original FWR MAV without spring was established. This was followed by the dynamic model and the characteristic equation of the system with a spring in two mounting positions in the mechanism. Based on the mathematical model, the average input power required for a flapping cycle of the system with and without spring was calculated as a measure of the efficiency of the mechanism. The analysis results show that spring of different stiffness leads to an efficiency increase to a different extent. In the case of adding a spring of stiffness k=1 N/mm, the input power for the system flapping at frequency 9 Hz was reduced by 54.5%. Secondly, in order to validate the theoretical model and results, a numerical simulation of the FWR model was also carried out by using the commercial software ADAMS. For different models, the maximum difference of the power results from the two methods is only 4.5%. Thirdly, in order to find the most efficient system, an optimization process was carried out for searching the minimum average power or the input torque peak value with the spring stiffness and length as design variables. The results show that the required input power for the optimized system was reduced by 63.0%. Finally, the experiment of the FWR test model was conducted to measure the flapping motion and the total force based on which the spring’s contribution to the average power reduction was calculated. Take the spring stiffness k=1 N/mm for example, the average power of the model flapping at a frequency 9 Hz can be reduced by 55.5%. This is very close to the results obtained from the theoretical and ADAMS models.