Gust generator
Introduction
In addition, I investigate the impact of various pitching maneeuvers on mitigating the lift impact of the gust. These maneuvers were based on open-loop and closed-loop strategies. Open loop strategies refer to maneuvers prescribed a-priori of the encounter based on full knowledge of the flow. Closed-loop strategy refers to a maneuver that is generated in real-time based on sensor measurements without knowledge of the flow. The video below shows the flowfield and force results from the closed-loop control maneuvers
Design
The water exits the pump's discharge side and is guided to the diffuser section by PVC piping. The diffuser sections expand the flow gradually to avoid the onset of flow separation and the formation of recirculation zones. The fours diffusers merge into the mixing chamber that creates some travel distance for the flow to undergo before further treatment. This gives enough time for the flows to merge. The flow then turns 90 degrees via the aid of turning vanes into the gust outlet. The flow passes through a section of 1/8" honeycomb flow straightener on its way out of the system to remove flow tilt as well as break down large turbulent structures.
The Diffuser-pipe adapter is a section that provides the flow a smooth transition from a circular pipe cross-section to a rectangular diffuser cross-section.
The diffuser section expands the flow gradually to increase the gust volume region. The diffuser has a maximum equivalent cone angle of 6° to avoid the onset of flow separation at the walls. This ensures a smooth uniform expansion of the flow. The diffuser length is chosen such that the the area expansion ratio is limited to 3 limiting the probability of flow separation.
The adapters were 3D printed using and FDM printer and PLA plastic. The diffuser sections we built out of cut HDPE plastic panels due to their cheap cost, good marine profile, and smooth surface. 3M 5200 marine adhesive seealant was used to seal the diffuser section seams. Epoxy resin was use to join the adpater and diffuser section. The picture on the right is a top view of the adapter and diffuser section.
The mixing chamber and the gust outlet were built out of Delrin plastic sheets. Delrin was chosen because it is easy to machine, has a smooth surface, and is naturally matte black which would minimize laser sheet reflection problems during Particle Image Velocimtery (PIV) experiments. The diffuser sections were not built of the same material because it is expensive and thus unafforadable for larger sections. The plastic sheets were joined using epoxy resin.
A 1" inch" thick, 1/8" cell Aluminum honey comb sheet is mounted at the end of the gust outlet to straighten the flow. The length to cell diameter ratio of the honeycomb cell is 8 which ensures proper flow straightening. The honey comb also serves the purpose of breaking down any remaining larger scale turbulent structures. Usually, a finer screen mesh section is introcued after the honey comb to break smaller turbulent structures and make the flow more uniform. However, the pressure drop introduced by the screen is too large and would lead to unaccaptably slow gust exit velocity for the water pump we have.
The turning manifold utilizes turning vanes to gently turn the flow 90° to create the tranverse gust. The turning vanes consist of 90° arcs with a large, 1" radius to minimze the risk of flow separation as the flow turns. tangents at either ends of the vanes are added to ensure the flow enter and exit the turns at the right angles. Turning vanes without these tangents were tested and the flow exiting the outlet was found to be tilted.
The chord to vane spacing ratio is 25% which has been determined by Salter* to produce a loss coefficient as low as 0.06. The turning manifold was printed using an FDM 3D printer and black PLA. The black PLA is chosen to minimize laser sheet reflection problems.
Salter, C., "Experiments on Thin Turning Vanes," ARC R&M 2469, Oct. 1946.
At the outlet, the flow passes through a section of honeycomb as well as two consecutive screens. The honeycomb section straightens the flow and breaks down any turbulent eddies larger than the individual honeycomb cell size. The honeycomb used has a cell length-to-width ratio of 8]. A coarse screen followed by a fine screen are placed after the honeycomb to obtain a spatially uniform flow across the outlet area.
Testing
Two-dimensional Particle Image Velocimetry (PIV) was used to obtain flowfield measurements. The PIV setup is illustrated in this figure. Flowfield data was recorded in the lab frame 1 chord away from the wing’s midspan. A Quantel Evergreen Nd:YAG 532 nm laser creates dual beams pulsing at a frequency of 15 Hz. Due to the low pulsing frequency of the laser, a single time-resolved time-series is not possible. However, the dual laser allows for the acquisition of 15 Hz image pairs with a small enough time between the single images within a particular pair to allow for accurate calculation of particle displacement. The time separation between images within a pair is picked such that the maximum particle displacement of particles across an image pair is 4-5 pixels. The laser beam is reflected via a set of 90 degree mirrors and transformed into a laser sheet using a 7 degree Powell lens. The laser sheet illuminates neutrally buoyant soda lime glass particles that accurately track the flow. The particle data is collected by a high-speed Phantom v641 camera with a 2560 x 1600 resolution.
The two pictures above show laser sheets in the spanwise (left) and chordwise (right) planes. Refer to the full CAD picture for the definition of these directions. These two sheets were used to obtain PIV flowfield measurments in the orthogonal planes to fully characterize the output of the gust outlet.
750 image pairs were acquired at 15 Hz for each case leading to a total acquisition time period of 50s per case. The flowfield was averaged over the acquisition period and the characteristic velocity profile of the gust was determined. The figures above show the averaged vorticity and velocity vectorfield data and the instantaneous data for the chordwise and spanwise directions. The left column shows the data for the spanwise direction and the right column shows the data for the chordwise direction. The velocity within the gust is uniform within the regions spanned by the gust’s outlet. The edges of the gust exhibit strong shear layers separating the gust flow from the surrounding quiescent fluid. While the average flowfield exhibit a smooth shear layers at the edge of the gust, the instantaneous shear layers exhibit some degree of unsteadiness and vortex shedding.
The top figure shows the time-averaged transverse velocity sampled along a horizontal slice at the travel height of the wing for different pump speeds. Figure a shows the chordwise velocity distribution which can be fitted to a trapezoidal trend. In an ideal world, the the gust would have demonstrated a top-hat velocity distribution. However, as the gust travels vertically, the shear layers diffuse out, and thus the velocity rise slope is dampened. Figure b shows the velocity distribution along the spanwise direction for different pump speeds. The two vertical black lines demarcate the gust region spanned by the wing. Although the transverse velocity exhibits a reduction towards the edges of the outlet in the spanwise direction, the tranverse velocity in the region spanned by the wing is consistent. The bottom figure shows the relationship between the pump speed and gust velocity. This linear relationship will be used to achieve gust velocities outside of the narrow range used to characterize the system.