- 30 views
- 33 downloads
Developing a GPU based Real-Time Particle Image Velocimetry System for Active Flow Control
-
- Author / Creator
- Friesen, Ivy O
-
The objective of this work was to create a Graphical Processing Unit (GPU)-based Real-Time Particle Image Velocimetry (RT-PIV) system for use in Active Flow Control (AFC) of a Turbulent Boundary Layer (TBL), and to evaluate the effect of effect of the flow control on the boundary layer. The experiment involved using the Cross-Correlation (CC) method utilizing Fast Fourier Transform (FFT) to compute particle displacement between two image pairs. The FFT method was chosen for its low computational cost, and ability to implement calculations on the GPU. Computations were done in MATLAB using a NVIDIA GeForce GTX 1080 Ti Graphical Processing Unit (GPU). The calculated RT-PIV vector fields were size 3 × 30, with 3 vectors in the streamwise direction and 30 vectors in the wall-normal direction. The average time for the RT-PIV code to process on the GPU was found to be 1.20 ms, with a standard deviation of 0.07 ms. The RTPIV AFC system was tested at 25 Hz in the University of Alberta wind tunnel. The time limiting factor for this test was the laser, the New Wave Research Gemini 30 Hz PIV Nd:YAG Laser system. A voice-coil actuator was used as the active surface actuator for flow control. This actuator was mounted flush to the floor of the wind tunnel and had a range of motion of ± 3 mm displacement. One streamwise vector from each vector field was used as the input to the controller. The calculated particle velocity between two frames was filtered using a first order Butterworth filter and a proportional control law was used to determine the signal sent to the actuator. This filter was used to attenuate higher frequency noise from the measured velocity. A parametric study was done on varying the filter cutoff frequency (5 Hz or 10 Hz) of the lowpass Butterworth filter, as well as the location of the vector used for flow control (8 mm or 11 mm above the wind tunnel floor). The Reynolds shear stress plots for each of the four experiment iterations were plotted and compared with and without actuation. The trial using a filter cutoff frequency of 5 Hz and input vector 11 mm from the wind tunnel floor showed the most promising results with a decrease in Reynolds shear stress when actuation was occurring, however, more tests are needed to ensure these results are statistically significant. Future work includes using a higher frequency, higher powered laser for better PIV image results, using a NVIDIA RTX 3090 GPU for faster algorithm speed, using an actuator with higher range of motion, and utilizing different control laws.
-
- Graduation date
- Fall 2024
-
- Type of Item
- Thesis
-
- Degree
- Master of Science
-
- License
- This thesis is made available by the University of Alberta Library with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.