The experiments were conducted in a channel with square cross-section (5 cm × 5 cm), 100 cm long. Air and water were used as the gas and liquid mediums, respectively. The channel was made of acrylic. A honeycomb was placed near the upstream end of the channel to straighten the flow and remove flow disturbances. Water was circulated through the channel via a magnetic pump (Little Giant, 5 MD) from a 60 gallon reservoir (see Figure 2-1). Water flow rates were adjusted using a rotameter (FP 1-35-G-10/83, F&P Co) which was installed downstream of the pump. The pressurized air was injected into the liquid stream in the channel via a nozzle to generate bubbles. To maintain steady supply of air, the compressed air from the main supply line was first passed through a
0.16 m3 tank, which served as a settling chamber, to remove pressure fluctuation. The air
from the settling chamber then passed through a needle valve into a long capillary tube that was connected to the nozzle. The needle valve was used to control the air flow rate, which was measured by a rotameter located upstream of the needle valve (see Figure 2- 1). The nozzle was located 70 cm downstream of the channel inlet. A prior set of experiments in the channel using Particle Image Velocimetry (PIV) technique confirmed that the channel flow was fully developed at the nozzle location. The uncertainties in
liquid and gas flow rates based on the rotameters used were ±3 and ±0.08 cm3/s.
As mentioned earlier, a novel nozzle design developed by Gadallah and Siddiqui [25] was used in this study. The unique feature of this nozzle design was the presence of side holes near the main nozzle rim, which generates small bubbles with higher detachment frequency. Two side-holes nozzles are referred to as Configuration A. Two orientations of the novel nozzle were considered in the study. In the first orientation (I), the side holes
were aligned with the liquid flow direction and in the second orientation (II), the nozzle
was rotated by 90o thus, the side holes were located perpendicular to the liquid flow
direction, hereinafter referred to as Configuration A-I and A-II (see Figure 2-2). A standard nozzle was also considered which served as a reference case. Both novel and standard nozzles have inner diameter of 0.82 mm and outer diameter of 1.62 mm and were made from brass tube. Three diameters of the side-holes were considered in this study, which were 0.5, 0.7 and 0.82 mm. Experiments were conducted at four liquid flow rates and three gas flow rates for each nozzle. Table 2-1 summarizes the gas and liquid flow rates and the corresponding velocities.
A high-speed camera (Photron SA5) with a 60 mm lens was used to capture bubble image. The camera has the resolution of 1000 × 1000 pixels up to 7500 frames per second. The camera resolution decreases with a further increase in the frame rate. The camera was connected to a PC and was operated via Photron FASTCAM Viewer software. The camera has a built-in memory card that allowed direct image recording. These images were later transferred to the hard drive. Back-lit shadowgraphy technique was used to illuminate the background for bubble identification. For this purpose, a 500 W halogen lamp was placed behind a diffusion screen to generate a uniform light. The images were captured at a rate of 1000 frames per second, and 2500 images were acquired and processed for each case. An in-house Matlab algorithm developed by Siddiqui and Chishty [19] was used to detect bubbles and to quantify various bubble parameters. Once a bubble is detached from the nozzle, the code automatically detects and tracks the bubbles and computes various bubble characteristics such as the bubble trajectory, detachment frequency, velocity, cross-sectional area and equivalent diameter. The uncertainty of detecting the bubble boundaries was within ±2 pixels, which correspond to the uncertainty of ±0.06 mm that translated into bubble diameter uncertainly of ±0.13 mm.
The experimental procedure used in this study is described as follows. The first set of experiments was conducted for stagnant liquid and varied the gas flow rates from 0.168
to 0.522 cm3/s, as mentioned earlier (see Table 2-1). In the following sets of experiments,
from a minimum to a maximum value. At each condition, the image acquisition started 10 minutes after setting the gas flow rate to reach the steady state.
(b)
Figure 2-1: Experimental setup; (a) Schematic and (b) photograph.
Table 2-1: Selected gas and liquid flow rates and corresponding average velocities
Water flow rates (cm3/s) 0 255 395 535
Water velocities (cm/s) 0 9.85 15.25 20.65
Gas flow rates (cm3/s) 0.168 0.280 0.522
Figure 2-2: Nozzle designs and orientations with respect to the flow direction used in the experiments. (a) Standard nozzle, (b) Novel nozzle, 2 side-holes in-line orientation, (c) Novel nozzle, 2 side-holes perpendicular orientation. hs=1.6mm.
To visualize the bubble formation and detachment mechanism and to study the gas behavior inside and outside the nozzle during bubble formation, a second set of the experiments was conducted using glass nozzles. Both standard and novel nozzles were made from glass tubes with inner and outer diameters of 0.99 mm and 1.28 mm, respectively. The novel nozzle had a side-hole diameter of 0.86 mm. Due to the difficulty in exactly matching the diameters of commercially available brass and glass tubes, the glass nozzles has slightly different dimensions compared to the brass tubes. However, they served the purpose of visualizing the underlying phenomenon. For the glass nozzle study, the experiments were conducted at the same liquid flow rates described earlier and
one gas flow rate (0.881 cm3/s). To accurately capture the underlying phenomenon, the
frame rate for the glass nozzle experiments was set at 10,000 frames per second. At each condition, the image acquisition started 20 minutes after setting the liquid flow rate to reach the steady state. For each case, 5,000 images were acquired.