Experimental investigation of flexibility effects in flapping wing aerodynamics

In the present PhD thesis, flexibility effects in flapping wing aerodynamics are experimentally investigated. For that purpose, one rigid and three chordwise flexible wings are produced and their aerodynamic performances are compared in a water channel. Digital Particle Image Velocimetry (DPIV) system is used to determine instantaneous velocity fields and vortical structures around the wings in conjunction with simultaneous direct force/moment measurements. The imaging system also acquired the instantaneous deformation of the wing. Natural frequency effects and deformation of the flexible wings are linked to the force measurements and efficiencies. Since all natural flyers and swimmers have some degree of flexibility on their wings, it is important to understand the phenomena for a more effective aerial/nautical vehicle design. Main objective is to determine the aerodynamic performances of different flexible wings under harmonic plunging or rapid pitch-up motions. The experiments are performed in the close-circuit, free-surface, large-scale water channel located in the Trisonic Laboratories at the Faculty of Aeronautics and Astronautics of Istanbul Technical University (Istanbul, Turkey). Two distinct motions are executed in the experiments. Perching or pitch-up is one of these motions and based on sudden change in angle of attack and halting the wing at a constant degree of angle. The pitch-up motions of the airfoil were accomplished using a Kollmorgen/Danaher Motion AKM33E servo motor. The models were subjected to performing two types of pitch-up motion: fast and slow. The model started from 0° and attained its final angle of attack of 45° in 1 second for the fast motion and in 6 seconds for the slow motion, corresponding to 1 and 6 convective times, respectively. The motor motion profiles were generated by a signal generator Labview VI (Virtual Instrument) for the given amplitude and duration. Labview VI also triggers both the force data acquisition and the PIV system. The other motion type is the purely plunging motion that is generated by harmonic cosine position function. Plunge motion of the model is accomplished with Kollmorgen/ Danaher Motion AKM54K servo motor. The motor motion profiles were generated by a Labview VI for various amplitudes and frequencies. The experiments are performed on four different wing materials: a rigid wing produced of plexiglass, three flexible wings produced of homogeneous polycarbonate sheet (lexan), an acetate wing, and an elastic model produced in 3D printer. The rigid wing has a thickness of 5 mm, chord length of 100 mm and transparent to illuminate both sides of the wing. Flexible wings have 10 mm metal part on leading edge to avoid spanwise flexibility and the total chord length is 100 mm. The wing's rotation axis is the leading edge, which also coincides with the center of the force sensor. Natural frequencies of the manufactured wings are one of the backbone parameters of the thesis. Motion parameters are determined based on these natural frequencies. The wings are exposed to sudden loading and given free chordwise oscillation to determine their natural frequencies while the wing is in its position, fully submerged in the water channel. Natural frequencies are determined as 7.40 Hz, 2.32 Hz, 1.25 Hz, 0.51 Hz for rigid, lexan, printed and acetate wings respectively. Force data is collected by a six-component ATI NANO-17 IP68 Force/Torque sensor. The sensor was mounted on the rod between the model and the plunge servo motor, oriented with its cylindrical z-axis coincides with the leading edge of the wings. For the perching rigid wings, whether the plate pitches up fast or slow, the flow remains mostly 2D along the span for plates with AR = 4 and above. During the motion, the effect of the AR is to shift the force variation in amplitude without changing its variation character. The effect of AR, hence, the three-dimensionality of the flow becomes pronounced after the motion ends. The formation of a local maximum in the force variations after the motion ends was evident for wings with AR > 2. There exist two peaks for wings with AR ≥ 6. These secondary or even tertiary peaks occur because of the formation of a new LEV. Even if the effective ARs of the wings are the same, the one with a symmetry plane did not yield exactly the same flow topology and hence the same force-time histories as the fully submerged wing. There were slight differences due to the three-dimensionality of the flow. Therefore, an endplate representing the symmetry plane should carefully be placed with special care to ensure that the gap between the wing root and the endplate is approximately 2 mm (i.e., 2% of the chord). Perching motion in flexible wings primarily depend on the deflection of the wings caused by the oncoming freestream. The deflection is high for the most flexible acetate wing, which results in a lower angle of attack. As a result, low CD is obtained for the acetate wing. When flow structures are examined, streamline topology of the acetate wing shows differences due to the reduced angle of attack both in fast and slow motion cases. Continuous LEV and TEV vortex shedding can be observed on the acetate wing. For the plunging rigid wings, it is found that the phase lag between the motion and the cross-stream force component decreases as the reduced frequency increases, and for a given reduced frequency, the phase lag increases as the plunging amplitude increases. The variation of the horizontal spacing between vortex cores with the reduced frequency follows the same characteristics as those of the phase lag for the cross-stream force. On the other hand, the phase lag for the streamwise force component decreases as the plunging amplitude increases for low reduced frequencies. The decrease converges to a phase lag value of 90°, which coincides with the maximum velocity of the wing. After the minimum is reached, the phase lag for the streamwise force component starts to increase. The minimum phase lag coincides with the reduced frequency where the crossover from drag to thrust is observed and maximum flapping efficiency is obtained. When the thrust is produced, the phase lag for the streamwise force component increases. Moreover, according to the vortex structures, the crossover from drag to thrust is observed when the LEV starts to hold around the leading edge and does not convect downstream along the plate. In flexible flapping wings, trailing edge amplitude makes a peak when the oscillations are performed near natural frequency of the wing (f/fn ~1). For the acetate wing, frequencies beyond natural frequency can be tested and after f/fn = 2, trailing edge amplitudes start to increase again. The values are in a continuous increase in the measurement range up to f/fn = 4.4. This finding suggests that, higher trailing edge amplitudes can be reached when the flapping frequencies are beyond the resonant frequency. When Φy phases are considered, it is observed that the flexible wings have completely different phase variations compared to the rigid wings. Natural frequency appears to have an effect on Φy phase variations. Phase values increase up to the near natural frequency, and then exhibit a decreasing followed by an increasing trend at higher flapping frequencies (can be observed for acetate wing). Besides the natural frequency, flapping amplitude also has an effect on phases such as the peak near natural frequency diminishes for higher amplitudes in acetate wing. This implies that natural frequency and flapping amplitude should be interpreted together. An interesting result is obtained from the combined Φy - ΦTE phases graph: trailing edge phases and Φy force phases are intersecting at different plunging frequencies for different wings. The intersecting frequencies coincide with the natural frequency of the wing. In other words, cross-stream force phases and trailing edge phases are equal when the wings are plunged near their natural frequencies. Flexible wings have peak efficiency around natural frequencies where their trailing edge phases and Fy force phases also coincide. Unlike the rigid wing, distribution of the Φx phases have similar characteristics with the Φy phases in flexible wings. That similarity indicates that the flexibility effects are observed not only in cross-stream direction but also in streamwise direction. In flexible wings, elastic forces emerge in both streamwise and cross-stream directions coupled with the trailing edge deflection. The similarity between the streamwise and the cross-stream force phases is a result of these prevailing elastic forces over the other existing forces (circulatory and added mass). The efficiency peaks shift towards high Strouhal numbers when the flapping amplitude increases. The peak efficiencies are obtained in a Strouhal number range of 0.1 – 0.6 for the flexible wings. High efficiency zone in the literature was expressed in a Strouhal number range of 0.2 – 0.4 and for the flexible wings, this range is generally valid. The boundaries of the range are slightly extended for the flexible wings in different flapping amplitudes. Same vortical instances are attained at different stages of the motion cycle for the flexible wings. It is observed that the TE motion determines the vortex shedding timing. As the wing gets more flexible, vortex shedding from the TE delays and similar vortex structures on the TE are observed later in the flapping period. Chordwise flexibility under parameters presented herein, alters the TEV shedding timing and vortex strength but does not have a pronounced effect on LEV shedding. For the rigid wing, the vortices burst after one chord length in the wake. As the wing gets more flexible, the vortices start to burst late. The vortices in the most flexible acetate wing's wake preserve their coherency and can be tracked in full measurement region. Also number of vortices observed in the wake increases when the wing is more flexible. While the flexibility for the wing flapping around its natural frequency exploits and enhances the high velocity zones created around the wing with the TE timing, another low efficiency flapping frequency cases cannot perform that. It can be considered as "catching" the flow and transferring it to the wake through vortices, which is found to be crucial for high efficiency.