Çırpan kanat aerodinamiği uygulamaları için, parçacık izleyerek hız belirleme yöntemi ile elde edilen verilerden kuvvet tahmini

The current study aims to develop an in-house code to estimate loads for a flapping wing using Digital Particle Image Velocimetry (DPIV) data. For the sake of simplicity, pure plunge motion is considered and experiments were performed using different space and time resolutions to assess the performance of the load estimation. Then, the developed code has been tested for different plunge motion cases. Although the study is focused on the load estimation, it make uses of a large quantity of DPIV and direct force measurement data. In order to calculate the force acting on the flapping airfoil, Reynolds Transport Theorem (RTT) can be utilized. As the velocity field is provided by the DPIV data in time, in order to compute the forces acting on the body, one has to obtain the pressure field around it. Present study includes the estimation of forces by using the pressure field which is obtained by the explicit integration of planar pressure gradients. The pressure gradients are obtained by the use of Navier-Stokes Equations in two-dimensional form. In Navier-Stokes Equations, material derivative is required to compute the pressure gradients. In order to compute the material derivative, two methods are used succesively in this work, namely Eulerian and Lagrangian approaches. The thesis describes the algorithm in detail with its variations. As a result higher order differentiation and surface integral for unsteady term yielded a better performance in force prediction. According to the sensitivity analyses, a spatial resolution of 3% of the chord and a field of view covering all the amplitude of motion gave a satisfactory results with a temporal resolution expressed as 40 vector fields per cycle of motion. In terms of DPIV post-processing, average correlation was found to reduce the noise as it is expected. On the other hand, the estimated load matched better the direct measurement when it was phase averaged and filtered. It should be noted that the compared direct measurement results were also 1 Hz filtered. For all the investigated results, both for the sensitivity analyses and general plunge cases, lift coefficient variation was in general predicted well, in agreement with the direct force measurements, however the estimated drag coefficient variation had a very large discrepancy with the direct force measurements. This might be due to the neglected viscous term. Consequently, the fact disabled thrust/drag analyses for plunge cases. Estimation of the lift force coefficient variation exhibited a nearly perfect match even on the instantaneous results when the plunge motion has the lowest frequency and the largest amplitude.