Air-conditioning is essential for keeping passengers comfortable aboard large-body aircraft. In the Airbus A330/A340 airplane, developers needed a maximum air-flow velocity of 0.25 m/s in the seating area, a level they reached by using computational fluid dynamics in the design of the cabin air-outlet sizes and locations.
After the design of the air-conditioning system was finalized, developers at Daimler-Benz Aerospace (Hamburg, Germany) turned to particle-image-velocity techniques to verify the design. To obtain quantitative results for the large and complex shaped cabin, they first filled soap bubbles with a mixture of air and helium. Then, they filled the rear part of the cylindrical fuselage section of an Airbus A330/A340 prototype with this mixture. To trace and understand the dynamics of the air flow, the bubbles were tracked and measured using picCOLOR, an image-processing system from The FIBUS Research Institute (Düsseldorf, Germany).
Installed in a PC system, the picCOLOR uses a Hitachi KPM1 768 × 512-pixel CCD camera that is coupled to an F64 PC-based frame grabber/image processor from Coreco (St Laurent, Quebec, Canada). Using 64 Mbytes of on-board DRAM, Daimler-Benz engineers were able to digitize 7 s of real-time or 14 s of half-speed image-bubble-motion sequences.
Because the amount of bubbles was not large enough to use frequency-domain-based correlation techniques, a fast particle-tracking velocimetry algorithm was developed by the FIBUS Research Institute engineers. After performing background removal and binarization, the imaging system calculated the bubble locations to sub-pixel accuracy, and then it determined the bubble velocity and the local flow velocity from the displacement of the bubbles in subsequent images.
Next, the generated randomly scattered velocity vector fields were interpolated to produce equally spaced grid vector plots and vorticity contour plots. Such plots showed that velocities of 0.05 m/s existed over the window seats and 0.1 to 0.25 m/s existed in all other areas of the aircraft. These data demonstrated excellent agreement with the results from the computational fluid-dynamics program.