PSI - Issue 5

Pedro J. Sousa et al. / Procedia Structural Integrity 5 (2017) 1253–1259 Pedro J. Sousa et al. / Structural Integrity Procedia 00 (2017) 000 – 000

1254

2

analysis (CFD), the simulation overestimates the velocity magnitudes in the rotor slipstream and presents discrepancies in the vorticity field [2]. These differences have been attributed to the lack of fluid-structure interaction due to deformation of the blades. The deformations change the effective blade pitch, resulting in rotor operating points distinct from their nominal values. The availability of blade deformation data can therefore enhance the quality of these systems’ numerical models. Measurement of blade deformations in operation is therefore gaining widespread attention. Typical analyses of these structures usually involve discrete sensors assembled on rotating blades and signal transmission through slip-rings [1, 3] or indirect measurement of the structural health [4]. Recently, image-based methods such as Digital Image Correlation have been successfully used for rotating targets [3, 5-11], enabling full field deformation measurements with a non-contact procedure under active research that has been gaining acceptance as a robust methodology in various fields [12-17]. The two main approaches for image acquisition of rotating targets are: achieving apparent stillness of the rotating target [3, 5-7, 11, 18-20]; and frame realignment using digital image processing methodologies [8-10]. Even though the latter could be used for rotor blades, it is not easily applicable to other rotating structures, such as shafts. It may be difficult to acquire images of the entire region of interest at the same time and the images ’ time reference is harder to define. Among the former, one option is the synchronized rotation of another object [18, 19], but it requires a precise alignment of the axis of rotation and only supplies a view along this axis. Stroboscopic lighting is another approach [3, 5-7, 21-23] that has been successfully applied to multiple fields and allows the use of lower cost commercial equipment, while enabling the monitoring of both rotor blades and shafts, since it can also provide a lateral view of the cylindrical surface. However, in these cases, it has only been applied to low rotation speeds of rotor blades. The custom image acquisition system used by [11, 20, 24, 25] is compatible with the stroboscopic lighting approach and can reach 100 000 rpm. However, its features are comparable to those offered by commercial high-speed cameras, requiring however a PC-based frame grabber. A stroboscopic lighting approach was also used by [26] to monitor a rotating shaft. However, due to frame rate constrains, it involved the under-sampling of the DIC signal and a constant acquisition frequency (i.e. the cameras are triggered by a clock signal and not by the position of the subject). Most of these approaches deal with the acquisition of the deformation of the object and, in a few cases, its modal analysis. Besides these, in [27] a DIC methodology is used to calculate the torque in short shafts under rotation. However, it does not achieve apparent stillness of the target, instead using constant framerate and requiring multiple reference images in different angular positions. The current work uses static illumination and high-speed cameras as a first approach in order to achieve similar results to the usage of stroboscopic lighting. Nonetheless, the developed framework is capable of working with stroboscopic lighting, if available. 2.1. Experimental setup The experimental setup is shown in Figs . 1 and 2. The target specimen was an RC helicopter’s blade, actuated by the toy’s control system. The rotating blades interrupt a laser beam that is aligned with a photodetector. This generates an electric signal that is sent to the trigger controller, which was developed using a digital signal processor. This controller is configured through a computer-based user interface in regards to timing, counting, among others. In this particular experiment, the trigger controller was configured to generate a pulse for every two interruptions, that is, once for every full rotation of the rotor. This trigger pulse is then sent to two high-speed cameras. Due to the unavailability of two high-speed cameras of the same model, a Photron FASTCAM SA4 and a Photron FASTAM SA3 were used. This caused some issues, such as the difference in trigger reaction times and image quality, which had to be accounted for. 2. Methodology