27 Toward Active Control of Limit Cycle Oscillations in an Aeroelastic Wing. . . 207 Fig. 27.2 Fully assembled and exploded view of the variable frequency disturbance generator showing the major components and subassemblies used during construction via steel rods. The sliding rail allowed the wing to translate, or heave, perpendicular to the direction of flow, while bearings in the carriages allowed the wing to rotate about its midchord. The support structure to which the rails were attached was constructed around the wind tunnel section using T-slotted aluminum framing. No components of the support structure were attached to the wind tunnel structure to avoid any unwanted vibrations being transferred into the experimental apparatus. The elastic properties of the system were supplied by attaching linear extension springs to the mounting carriages. The extension springs providing stiffness in the heave degree of freedom were attached to either side of each carriage and a fixed point on the support structure. Springs providing stiffness in the pitch degree of freedom were mounted on the carriages themselves and attached to a pulley connected to the steel rod connecting the wing and carriage. This allowed the pitch springs to translate with the carriage as it moved along the linear mounting rail. In total, eight springs were used, four for each degree of freedom split between the upper and lower carriages. The variable frequency flow disturbance generator, shown in Fig. 27.2, was designed to produce a well-defined von Karman vortex street with shed vortices being produced at rates from less than 1 Hz to at least 8 Hz. The main body of the disturbance generator was constructed from a 10.48 cm diameter cylinder made of braided carbon fiber produced by DragonPlate™ (Elbridge, NY, USA). Research by Rockwood and Medina [19] showed that while an oscillating cylinder can produce a locked-in vortex wake with a shedding frequency equal to its oscillation frequency, the inclusion of a trailing edge splitter plate resulted in a more well-defined wake for the same range of oscillation frequencies. For this work, the splitter plate was constructed from a 1.5875 mm thick carbon fiber and birch composite material produced by DragonPlate™ (Elbridge, NY, USA), which extended one cylinder diameter from the trailing edge. The rotational axis of the disturbance generator was placed at the central axis of the primary cylinder. A SureServo SVL-210b produced by AutomationDirect (Cumming, GA, USA) was used to drive the rotation of the disturbance generator and controlled using a Copley Control (Canton, MA, USA) Xenus XTL-230-18. A maximum continuous torque of 3.3 Nm and maximum instantaneous torque of 9.9 Nm was provided by the servomotor. While the Xenus controller provided direct control of the servomotor, an analog voltage trajectory sinusoid was supplied to the Xenus by a Keysight (Santa Rosa, CA, USA) 33500B Waveform Generator with the output controlled by the main data acquisition and control virtual instrument (VI). National Instruments (NI) LabVIEW was run on a NI PXIe-1078 data acquisition computer and used to construct a VI to gather real-time data and control the output status of the disturbance generator trajectory. Wing pitch and disturbance generator angle were recorded with US Digital (Vancouver, WA, USA) E6-10000 optical encoders. The disturbance generator encoder was placed just above the servomotor mount and the wing encoders were place both above and below the wing and the average angle used in all calculations to mitigate effects from structural twist in the wing. Linear displacement of the wing was measured using a Renishaw (West Dundee, IL, USA) LM10 magnetic linear encoder. Control of the disturbance generator was performed entirely within the LabVIEW VI used to run the experiment and record data. There were two methods of control used in the course of this work. The first method, User Control, allowed the researchers to manually set the oscillation amplitude and frequency and to start and stop the disturbance generator oscillation at-will. The second method, Automatic LCO Control, used preset oscillation amplitude and frequency and was started and stopped using wing pitch amplitude thresholds, allowing this control scheme to function as a rudimentary state machine
RkJQdWJsaXNoZXIy MTMzNzEzMQ==