Flow Field - Direct Injecion for Cat-Heating Operating Points
Beschreibung
PIV was used to measure the flow fields and their velocities. A double-pulsed high-speed Nd:YVO4 laser (Edgewave IS4II-DE, 532 nm) was formed into a laser cut with several optics (250 µm in the laser cut thickness) and then coupled vertically into the combustion chamber via a 90° mirror through the Bowditch piston assembly and the quartz glass piston window in order to illuminate the symmetry plane there. In the combustion chamber itself, the silicone oil droplets (0.5 µm, Dow Corning 510, 50cST atomized by Palas AGF 10.0) used as seeding particles were illuminated. These were mixed with the incoming air. The Mie scattering of the oil droplets was recorded using a Phantom V2640 high-speed camera in double-frame mode with a Sigma lens (focal length 180 mm, f# 5.8). The curvature of the cylindrical glass was compensated with a cylindrical corrective lens (+2000 mm). A high-resolution calibration target (LaVision) and polynomial dewarping were used in post-processing to transform the image into a real coordinate system and remove residual distortions caused by the imaging optics and the 20 mm thick cylindrical glass.
For the high-speed images, two images are taken at very short intervals. The time interval Δt between the two PIV images was optimized for the respective °KW under investigation in order to achieve optimal particle displacement and minimize particle losses perpendicular to the light section. This enabled high correlation values. As part of the measurements carried out, the time interval Δt was increased from 2 μs in the inlet stroke to 20 μs in the direction of the compression OT.
The vector calculations were carried out using DaVis 10.2.1 software (LaVision). A multipass cross-correlation with decreasing window size (64 x 64 pixels, 50% overlap, 32 x 32 pixels, 50% overlap) was performed. The calculated vectors were further processed without smoothing. Vectors with a correlation ratio between the first and second peak below 1.3 were deleted and a universal outlier detection based on a normalized median test in a sliding window of 7 x 7 vectors was applied. The resulting high-resolution vector fields were further processed in Matlab (Mathworks). In Matlab, the areas of engine components, such as pistons or spark plugs, were masked and statistical analyses were carried out with the remaining flow fields.
PIV was used to measure the flow fields and their velocities. A double-pulsed high-speed Nd:YVO4 laser (Edgewave IS4II-DE, 532 nm) was formed into a laser cut with several optics (250 µm in the laser cut thickness) and then coupled vertically into the combustion chamber via a 90° mirror through the Bowditch piston assembly and the quartz glass piston window in order to illuminate the symmetry plane there. In the combustion chamber itself, the silicone oil droplets (0.5 µm, Dow Corning 510, 50cST atomized by Palas AGF 10.0) used as seeding particles were illuminated. These were mixed with the incoming air. The Mie scattering of the oil droplets was recorded using a Phantom V2640 high-speed camera in double-frame mode with a Sigma lens (focal length 180 mm, f# 5.8). The curvature of the cylindrical glass was compensated with a cylindrical corrective lens (+2000 mm). A high-resolution calibration target (LaVision) and polynomial dewarping were used in post-processing to transform the image into a real coordinate system and remove residual distortions caused by the imaging optics and the 20 mm thick cylindrical glass.
For the high-speed images, two images are taken at very short intervals. The time interval Δt between the two PIV images was optimized for the respective °KW under investigation in order to achieve optimal particle displacement and minimize particle losses perpendicular to the light section. This enabled high correlation values. As part of the measurements carried out, the time interval Δt was increased from 2 μs in the inlet stroke to 20 μs in the direction of the compression OT.
The vector calculations were carried out using DaVis 10.2.1 software (LaVision). A multipass cross-correlation with decreasing window size (64 x 64 pixels, 50% overlap, 32 x 32 pixels, 50% overlap) was performed. The calculated vectors were further processed without smoothing. Vectors with a correlation ratio between the first and second peak below 1.3 were deleted and a universal outlier detection based on a normalized median test in a sliding window of 7 x 7 vectors was applied. The resulting high-resolution vector fields were further processed in Matlab (Mathworks). In Matlab, the areas of engine components, such as pistons or spark plugs, were masked and statistical analyses were carried out with the remaining flow fields.
Sammlungen
Die folgenden Lizenzbestimmungen sind mit dieser Ressource verbunden:
Solange nicht anders angezeigt, wird die Lizenz wie folgt beschrieben: Creative Commons Attribution-NonCommercial 4.0