Test facility

All testing was completed in the National Research Council of Canada's (NRC) High Pressure Spray Facility (HPSF) in Ottawa, Canada. This test rig consists of a pressure vessel with large optical access that can be used to study sprays over ambient pressure conditions ranging from near atmospheric to 20 bar(a). The piping and instrumentation diagram used for all tests is shown in Figure 2. Process control is achieved via Lookout software, while operating conditions are monitored and recorded using LabView software. Further details on the operation and capacity of this facility can be found in (Corber, 2007).

Figure 2.

Piping and Instrumentation Diagram.

Optical diagnostics

In order to assess the capability of the SLIPI based LIF/Mie ratio-metric method to provide robust and quantitative droplet sizing data, experiments were carried out on sprays over a range of liquid flows and ambient pressures using both conventional and alternative fuels as well as a calibration fluid as shown in Table 1. Conventional laser sheet images were collected to help visualize global spray features as well as to offer a comparison to the SLIPI images. PDA measurements were made at selected points in order to calibrate the SLIPI images and obtain SMD maps of the sprays.

The conventional laser sheet imaging instrumentation employed a 5W DPSS laser, manufactured by UltraLasers, as the light source. It operates at 532 nm and was equipped with sheet forming optics. A Cannon 5D Mark II digital SLR, equipped with a Canon 24–105 f4.0 L-series lens, was used to capture the images. The ratio-metric imaging technique involves doping the fuel with a fluorescing dye, and illuminating the spray with a laser sheet. Two images of the spray are then collected simultaneously, with one of the cameras using a filter to collect the LIF signal, and the other the Mie signal. These two images are then combined to give the LIF/Mie ratio which is nominally proportional to SMD. A more detailed explanation of the technique can be found in (Mishra et al., 2015). The system used to capture LIF/Mie images uses the same laser as the conventional imaging that is then modulated using a universal grating based structured laser illumination planar imaging optics module manufactured by LaVision Inc. Two LaVision Imager ProX cameras controlled by a programmable timing unit are used to capture pictures of the spray. The system was set up to only view half the spray, which allowed for a greater vertical field of view, while still optimizing the resolution of the camera's chip. The configuration is shown in Figure 3.

Figure 3.

Experimental set up of LIF/Mie Ratio-metric imaging system. Top View.

For conventional drop sizing and velocity measurements, a commercial Phase Doppler Particle Analyzer (PDPA), manufactured by TSI Inc. was used in this study. The system consists of a TM250 transmitter, equipped with a beam expander and a 500 mm lens. This results in a velocity range of −225 m/s to 1,125 m/s. The receiver, model number RV2100, was orientated in the 30 degree forward-scatter configuration. Using this optical path, the PDPA can measure particle diameters from approximately 1–266 µm.

Fuel nozzle

All testing was conducting using an aero-engine derived fuel injector. A photograph of the nozzle is shown in Figure 4. For the results reported here, only the pilot fuel line was used, which is a simple-orifice type pressure atomizer. It has a flow number (FN) of 0.2 g/s/kPa^1/2 and produces a hollow cone spray. For further details on the design and operation of the fuel injector see (Corber et al., 2018).

Figure 4.

Photograph of the fuel injector. Primary fuel passage:simple-orifice pressure atomizer. Secondary fuel passage: air-blast atomizer (not used in this study).

Test liquids

Three test liquids were employed in this study. A brief summary of their distinguishing features is shown in Table 1. These fluids were selected for their range of physical properties, as well as represent separate categories: calibration fluid, conventional jet fuel, and alternative jet fuel respectively.

Test matrix

The complete test matrix for the experimental campaign is shown in Table 2. It is derived from the test points provided by the National Jet Fuel Combustion Program (NJFCP), more details of which can be found in (Colket et al., 2017). It has been expanded slightly to cover a wider range of flows than those used in previous work (Corber et al., 2018) so that the application of the SLIPI LIF/Mie method can be tested over a greater variety of spray shapes, sizes, and droplet diameters. All three fluids were tested at these operating conditions.

Conventional laser sheet imaging - results

A sample of the conventional laser sheet images collected is shown in Figures 5 and 6. These pictures offer qualitative details with regard to the spray's overall features such as cone angle and quality of atomization. Details with regard to the spray's interior structures, such as if the spray is a solid or hollow cone, cannot be determined from these images. At the operating conditions presented, very little difference is observed between the various fuels. This is consistent with data found in the litterature for this type of injector (Lefebvre and McDonell, 2017). The results demonstrate that the spray changes significantly with rig operating pressure. As the ambient pressure increases the spray's cone angle collapses, resulting in a much narrower spray.

Figure 5.

Conventional laser sheet images of MIL-C, A-2, C-3 fuel sprays. Fuel flow 5 g/s. P3 = 3.4 atm. Camera settings: 1/2000s, f/6.3, ISO-400.

Figure 6.

Conventional laser sheet images of MIL-C, A-2, C-3 fuel sprays. Fuel flow 5 g/s. P3 = 10 atm. Camera settings: 1/4000s, f/6.3, ISO-400.

Phase doppler anemometry results

Sample sets of PDPA data showing the radial distribution of droplet SMD for two test fuels at two axial locations downstream of the injector are given in Figures 7 and 8. During commissioning of the experimental setup, data points were taken across the entire width of the spray, and the measured droplet sizes were found to be symmetric about the centreline. Subsequent tests were then reduced to sampling on only half of the spray to lessen time and cost.

Figure 7.

Droplet SMD measured by PDA at 25.4 mm from the nozzle face for MIL-C, A-2 and C-3, where the filled and un-filled symbols represent the 10 atm and 3.4 atm conditions respectively. Fuel mass flow 5 g/s.

Figure 8.

Droplet SMD measured by PDA at 50.4 mm from the nozzle face for MIL-C, A-2 and C-3, where the filled and un-filled symbols represent the 10 atm and 3.4 atm conditions respectively. Fuel mass flow 5 g/s.

The PDPA was set up to collect 50,000 samples per point. A validation rate of 90% was achieved for the majority of cases, resulting in 45,000 valid droplet samples at each radial location.

The PDPA data qualitatively agree with the laser sheet images, indicating that a higher ambient pressure results in narrow sprays. It also indicates that the lower viscosity fuel, A-2, atomizes better (smaller droplets) than the higher viscosity C-3 fuel. This difference in atomization was however found to diminish at higher ambient pressures for the injector used in the investigation. Additionally, these plots indicate that the droplets size increases with increased ambient pressure as well as with increased axial position. These results agree well with those found in the literature (McDonell et al., 1994).

SLIPI ratio-metric images results

Figure 9a shows the typical resulting image generated by the SLIPI LIF/Mie ratio-technique. It is the result of the combination of 150 Mie and 150 LIF SLIPI images. The normalized counts displayed in the figure are effectively the pixel-by-pixel ratio of the LIF and Mie images. Provided the fluorescing dye is evenly distributed throughout the liquid, the LIF signal is roughly proportional to the droplet volume. Since the Mie signal is roughly proportional to the droplet surface area, higher counts indicate areas in the spray with larger droplets (i.e. larger volume to surface area). Based on the work of (Mishra et al., 2015), the LIF/Mie ratio pixel count should correspond to a discreet SMD anywhere in the image. However, the count value itself is dependent on a number of parameters including, but not limited to, laser power, camera exposure time, particle number density, dye concentration, as well as the image processing methods. As a result, the absolute value of the pixel count is arbitrary, so in order to compare one ratio-metric image to another, pixel counts should be normalized. These normalized pixel counts cannot be used to determine droplet size directly, so in order to convert the image into a map of the droplet sizes an in-situ calibration is required to convert the intensity ratios to SMD's. The droplet size data collected by PDA was used to accomplish this, with the white dots in Figure 9b showing the corresponding location of those measurements.

Figure 9.

Uncalibrated SLIPI LIF/Mie ratio-metric image of C-3 spray with geometric scale. Fuel flow 5 g/s, P3 = 10 atm. (a) Raw image. (b) Post processed image with location of PDPA data.

This technique does require a limited amount of image post processing. In order to reduce residual noise, a 20 × 20 pixel median filter, equivalent to 1 mm × 1 mm area in the flow, was applied to the images. Figure 9b shows a resulting post processed image.

PDA data was collected at two or more axial locations in the spray to determine the robustness of the calibration within a single image. This was also done to help determine if the location of the calibration measurements affects the SMD to pixel count calibration curve. To determine the consistency of the calibration across operating conditions and test fuels, calibration curves have been generated for MIL-C, A-2 and C-3 at two rig operating pressures for a liquid flow rate of 5 g/s. The results are shown in Figure 10.

Figure 10.

Calibration curves for MIL-C, A-2, C-3 LIF/Mie ratio-metric images at P3 = 3.4 atm and 10 atm. Fuel mass flow 5 g/s.

Figure 10 shows the overall shape of the calibration curves to be consistent across all cases. At the 10 atm condition the calibration curves generated from the PDA data at 25.4 mm and 50.8 mm planes correlate closely, while for the 3.4 atm condition, the correlation between the two axial planes is not as high. Of all cases shown, the C-3 fuel at 10 atm offers the best results, with a nearly identical pixel count to SMD curve. There are several factors that help explain this. Based on the PDA data, the C-3 fuel at the 10 atm condition produces the spray with widest range of droplet sizes. As a result, this offers the largest calibration range, helping to reduce uncertainty. In addition, these larger droplets fluoresce and scatter the light with greater intensity. This leads to superior signal to noise ratios which also assists in improving the accuracy.

In order to determine the exact mechanisms causing the difference between the upper and lower planes for the 3.4 atm condition (i.e. Figure 10a–c) further investigation is required. A detailed examination into the sources of uncertainty, in conjunction with a statistical analysis for a large number of cases will need to be completed before conclusions can be drawn.

A second order exponential decay function was fit to relate pixel count to SMD in the LIF/Mie images. When applied to the image shown in Figure 9b, the result is a planar map of the droplet SMD's in the spray, and is shown in Figure 11. This plot is generated from a single calibration curve. It is observed to have the same overall trend as the PDPA data with larger droplets at the exterior of the spray, which increase in size as they travel downstream. Based on Figure 11, a single SLIPI LIF/Mie ratio-metric image can provide a spatially resolved droplet SMD's map. This is significant, since in order to produce the same chart, hundreds of PDA points and several hours of testing would be needed.

Figure 11.

Planar SMD map (Calibrated SLIPI LIF/Mie ratio-metric image) of C-3 spray. Fuel flow 5 g/s, P3 = 10 atm.

Figure 12 shows a comparison between a conventional laser sheet image, and a calibrated SLIPI ratio-metric image. While both offer details about the shape of the spray, there is significantly more information in the latter. The SLIPI demonstrates the hollow cone nature of the spray, as well as trends in droplet size. Even though a calibration may still be required in many applications to determine the exact SMD distribution, these results show a measurable gain in capability over conventional diagnostics.

Figure 12.

Calibrated SLIPI LIF/Mie ratio-metric image of C-3 spray overlaid on conventional laser sheet image. Fuel flow 5 g/s, P3 = 10 atm.