Applied Physics B

, Volume 117, Issue 1, pp 401–410 | Cite as

20 kHz toluene planar laser-induced fluorescence imaging of a jet in nearly sonic crossflow

  • V. A. Miller
  • V. A. Troutman
  • M. G. Mungal
  • R. K. Hanson
Article

Abstract

This manuscript describes continuous, high-repetition-rate (20 kHz) toluene planar laser-induced fluorescence (PLIF) imaging in an expansion tube impulse flow facility. Cinematographic image sequences are acquired that visualize an underexpanded jet of hydrogen in Mach 0.9 crossflow, a practical flow configuration relevant to aerospace propulsion systems. The freestream gas is nitrogen seeded with toluene; toluene broadly absorbs and fluoresces in the ultraviolet, and the relatively high quantum yield of toluene produces large signals and high signal-to-noise ratios. Toluene is excited using a commercially available, frequency-quadrupled (266 nm), high-repetition-rate (20 kHz), pulsed (0.8–0.9 mJ per pulse), diode-pumped solid-state Nd:YAG laser, and fluorescence is imaged with a high-repetition-rate intensifier and CMOS camera. The resulting PLIF movie and image sequences are presented, visualizing the jet start-up process and the dynamics of the jet in crossflow; the freestream duration and a measure of freestream momentum flux steadiness are also inferred. This work demonstrates progress toward continuous PLIF imaging of practical flow systems in impulse facilities at kHz acquisition rates using practical, turn-key, high-speed laser and imaging systems.

1 Introduction

This work demonstrates the use of relatively practical, turn-key laser and imaging systems for tracer-based planar laser-induced fluorescence (PLIF) imaging of a practical flow field, an underexpanded jet in crossflow in an expansion tube. The transverse jet in crossflow (or JICF) is a basic configuration used to mix fuel (the injectant) with oxidizer (the crossflow) in an air-breathing engine. The entrainment, turbulent transport, and mixing characteristics of the jet in crossflow can play a major role in the performance of a combustion engine, and these characteristics are largely controlled by a complex system, of large-scale vortical structures [1, 2, 3, 4].

The jet in crossflow is particularly relevant to next generation, high-speed aerospace propulsion systems, such as a starting or unstarting scramjet (where the crossflow may be supersonic) or thrust vectoring (where the crossflow is subsonic). The specific configuration studied in this work, the jet in nearly sonic crossflow, resembles the fuel injection scheme of non-premixed rotary detonation engine [5]. In a non-premixed rotary detonation engine, fuel is injected into a subsonic crossflow or co-flow of air, and an unsteady detonation wave burns the mixture. This detonation wave also interacts with the jet of fuel and air, contributing to their mixing. Sufficient mixing of the fuel and air must occur before the next arrival of the detonation wave. In this work, the flow field generated in an expansion tube resembles the situation immediately behind the RDE detonation wave, but without the effects of chemistry, and high-repetition-rate (20 kHz) tracer-based PLIF is used to visualize the interaction of the shock wave and the jet.

Since its introduction, PLIF has enabled researchers to study a variety of flow fields and flames [6, 7, 8, 9]. PLIF imaging can utilize electronic or rovibrational [10] transitions in species that are naturally occurring in the flow field of interest, such as the hydroxyl radical (OH) or nitric oxide (NO) [7]; these experiments are convenient in that no additional species must be added to the flow field and insight can be gained into natural processes, such as reaction zone location and size [11, 12]. However, these naturally occurring LIF species typically require a tunable light source for excitation, and fluorescence quantum yields (FQY) are typically small and signal levels are correspondingly low.

As an alternative to imaging a naturally occurring species, a tracer molecule with favorable fluorescent properties can be added to the flow and imaged. Tracer-based PLIF can be made quantitative by using a tracer that has well-characterized photophysical properties, and then, the imaged LIF signal can be converted to physical quantities such as temperature or tracer mole fraction [13]. A tracer can also be selected that is particularly well suited for a given study, such as using toluene to image shock waves [14, 15], acetone to image the scalar mixing field in a jet in crossflow [2, 16], or NO to image high-temperature flows [17]. Tracers with a large FQY can also be utilized so that large fluorescence signals are generated, thereby improving image quality (i.e., signal-to-noise ratio, SNR). Many tracers also absorb broadly in the UV, and tracer-based PLIF can greatly simplify an experiment by allowing one to use a turn-key excimer or frequency-tripled or frequency-quadrupled Nd:YAG laser as the light source.

Recently available high-repetition-rate light sources and UV-sensitive cameras have enabled high-repetition-rate (i.e., >1 kHz) tracer-based PLIF to be used in the study of a variety of flow fields, such as biacetyl PLIF used to image mixture fraction in direct-injection engines at 12 kHz [18], acetone PLIF to image mixture fraction in a bench-top setting at 9.5 kHz [19], and toluene PLIF to image temperature stratification near walls (10 kHz [20]) and inside an internal combustion engine (6 kHz [21]). Toluene PLIF has also been used for single-shot imaging of supersonic flows generated in an impulse facility [22] and applied specifically to study the mixing of a jet in supersonic crossflow [23] in the context of scramjet propulsion.

Impulse flow facilities generate short-duration flows (\(\approx\)1 ms) and offer the capability to ground test the next generation of aerospace propulsion systems. A variety of laser-based diagnostics have been used to study flows generated in these facilities, but many of the techniques have not been capable of kHz acquisition rates until recently; Jiang et al. [17] provides an excellent overview of diagnostic techniques applied to flows generated in impulse facilities. The current study extends practical high-repetition-rate tracer-based imaging to an impulse flow facility. This extension may not seem particularly novel because this technique has already been demonstrated in applied settings such as internal combustion engines; however, due to the demanding nature of larger-scale flow facilities capable of generating high-speed or supersonic flows and the additional personnel and resources required to manage and operate these facilities, the turn-key nature of the laser and imaging systems used in this work creates new opportunities to use high-repetition-rate PLIF to study these practical flow configurations relevant to aerospace propulsion.

The short duration of flow fields typical of impulse facilities requires high acquisition rates (>1 kHz) to collect meaningful quantities of data in a reasonable amount of time. The development of CMOS cameras and high-speed intensifiers has increased image acquisition rates to more than 10 kHz [19], but conventional Nd:YAG or excimer lasers typically operate at low repetition rates (10–100 Hz), which has limited experimentalists to acquiring only a single image per shot in an impulse facility. Continuous-wave (CW) UV lasers have been proposed [24] and used [25] to increase acquisition rates, but the relatively low average power requires long integration times, and significant blurring or smearing in the image can occur for all but slow-moving flow fields.

Custom burst-mode lasers [26, 27] have been successfully designed and built for kHz rate PLIF imaging, and Jiang et al. [17] have used such a systems for NO PLIF imaging of a jet in crossflow within the 48-in. CUBRC shock tunnel. This manuscript is a complementary demonstration of high-repetition-rate PLIF in an impulse facility, but instead of using a custom-built burst-mode laser, an optical parametric oscillator, and imaging NO, the current work uses toluene as a tracer molecule and an easy-to-use, off-the-shelf, continuously pulsed 266-nm light source. A solid-state diode-pumped, pulsed Nd:YAG laser at 266 nm (delivering 0.8–0.9 mJ at 20 kHz for 16–18 W of average power) is used to excite toluene, and this laser can easily be operated single handedly. An intensified, CMOS imaging system collects the fluorescence. The tracer, toluene, is seeded into the freestream, and the lack of signal indicates the presence of injectant or absence of freestream fluid, enabling visualization of mixing, trajectory, and structure of the jet in crossflow. Each test lasts for approximately 18 ms, and 360 images are acquired for each test at 20 kHz. In addition to visualizing the dynamics of the jet itself, we also ascertain the steadiness and duration of the crossflow generated in the expansion tube. Collected images are presented in a movie, still frames, and averages; signal and SNR time-histories are also presented to provide a basis for design and implementation of this type of imaging scheme to study other flow fields within impulse facilities.

2 Experimental setup

A schematic of the experimental setup is provided in Fig. 1. An expansion tube serves as the impulse flow facility used to generate the flow field in this work. Expansion tubes [28] can generate a wide range of conditions by utilizing an unsteady expansion to process shock-treated gas to a final, test gas condition. Typically, expansion tubes are used to generate short bursts (<1 ms) of high-enthalpy gas with accurate freestream chemistry (i.e., low dissociation), but by changing gases and fill pressures, relatively long-duration (\(\approx\)1–2 ms) cold flows (\(\approx\)500 K) can also be produced [22]. An expansion tube can also be operated in ‘shock-mode’ by removing the secondary diaphragm to produce long-duration (\(\approx\)20 ms) flows behind the incident shock wave, which is how the tube is operated in this work.
Fig. 1

Schematics of the experimental facility and setup (not to scale)

In this shock-mode configuration, the incident shock wave propagates down the tube, heating, pressurizing, and accelerating the gas behind it, and this incident shock and the test gas flow terminate in a large-volume dump tank. The tube is outfitted with arrays of piezo-electric time-of-flight shock sensors (referred to as shock counters) used to measure the shock speeds and dictate the timing of data acquisition; the shock speeds are also used to infer the test gas conditions. For this work, the driver gas is argon (6.8 bar of pressure), and the driven gas is a mixture of helium and toluene (\(\approx\)5 % toluene by volume mixed manometrically in a separate mixing tank, 130 mbar fill pressure), resulting in relatively weak shocks and low post-shock temperatures. The resulting post-shock freestream pressure, temperature, and Mach number are 500 mbar, 480 K, and 0.9, respectively.

A flat plate (100 mm wide by 155 mm long) is mounted in the test section of the expansion tube; Fig. 1 includes a schematic of the plate, which is positioned at the exit of the expansion tube. A 2-mm-diameter contoured nozzle is machined into the plate 64 mm downstream from the leading edge. A regulated tank of hydrogen is connected to the plate via stainless plumbing, and a fast-acting solenoid valve (Valvetech model 15060-18) controls the flow of the jet fluid.

The imaging systems consists of a VisionResearch Phantom v710 CMOS camera and LaVision high-speed intensified relay optic with a Gen II photocathode, a ‘slow’ P46 phosphor, and a Sodern \(\frac{f}{2.8}\) 100 mm UV lens. A Schott glass 2-mm-thick WG280 long-pass filter is placed in front of the camera to block the majority of scattered 266 nm light; the filter transmission is roughly 1 % at 266 nm, 86 % at 300 nm, and >99 % from 300 nm to 2 μm; toluene fluorescence peaks near 280 nm and extends out to about 340 nm [13]. The phosphor in the intensifier has a relatively slow decay, and significant residual signal, roughly 6 %, is observed in the first subsequent image at an acquisition rate of 20 kHz. The residual signal due to the slow phosphor is removed using a walk-forward algorithm detailed in ‘Appendix’.

The light source is an EdgeWave HD40II-E Nd:YAG laser delivering 6 ns pulses of 266 nm light. The laser continuously outputs between 0.8 and 0.9 mJ per pulse, corresponding to an average power between 16 and 18 W. A beam-splitter, photodiode, UG11 filter, and integrating sphere are used to monitor shot-to-shot variation in laser energy and timing of the laser and intensifier gate. In tracer-based PLIF experiments, the windows of pressure vessels or static cells can be irreparably damaged due to the adsorption and subsequent burning of tracer if the windows are exposed to high average power for long periods of time; for this work, due to the extremely high average power, windows burn in just a couple of seconds if containing a mixture of test gas, so a mechanical shutter (Electro-Optical Products Corporation model SH-20) blocks the beam during the majority of the time to prevent damage to the test section windows (UV-grade fused silica).

For all images acquired in this work, gate times are 1000 ns, and the laser pulse is positioned (in time) 100 ns after the opening of the intensifier gate; fluorescence lifetimes of toluene are well under 100 ns [29]. The image size is 592 × 512 (\(W \times H\)) pixels, resulting in a spatial resolution of 10.5 pixels per mm or 95 μm per pixel. A 400-mm plano-convex cylindrical lens focuses the beam into a sheet, which is approximately 700 μm thick at full-width half-max and 1.75 cm wide. Two separate fields of view are imaged and then stitched together to form the full field of view visible in the presented images. The different fields of view are imaged during different expansion tube shots by moving the final reflecting and focusing optics downstream 1.25 cm between runs. Five millimeters of overlap is maintained between the two fields of view. Only the light sheet is moved to image the different fields of view, and so the camera stays in the same position resulting in simple image registration between upstream and downstream views. Roughly half of the laser energy per pulse is lost across the first optic due to the drop in reflectivity induced by the high temperatures created by the incident beam; mirrors downstream of the first optic perform near their specification (\(\approx\)99 % reflectivity). In the field of view, the laser fluence is roughly 3.5 \(\frac{\hbox {mJ}}{\hbox {cm}^2},\) which is below the fluorescence saturation level for toluene [29, 30].

Corrections are made to all images in the following process: first, all images are dark noise subtracted (50-image averages), and then, signals from slow phosphor decay are removed (details in ‘Appendix’). Then, background signals resulting from reflections in the confined environment (i.e., test section) are corrected for using the procedure outlined by Miller [31]. PLIF imaging in confined environments can result in large background signals due to the reflection of fluorescence off of surfaces (e.g., the back wall), especially when the freestream is seeded with tracer (opposed to, for example, OH imaging, where signal may only be collected from thin reaction zones). Another technique for correcting for these background signals is via structured light illumination [32, 33], which is particularly effective when using PLIF to image flow fields containing sprays or droplets in which the scattering or reflecting media are constantly changing. Miller [31] presents a model and method for correcting for these background signals based on deconvolution of the acquired image with a background kernel, which is computed using the dimensions of the confining environment. This kernel-based correction technique is useful for high-repetition-rate imaging of dynamic flow fields with static background (e.g., no particles or droplets) because it can be applied entirely during post-processing without the need for additional hardware or simultaneous acquisition of structured light images. In this work, background signals of up to 15 % of the maximum signal in the image are observed and successfully removed via the kernel-based correction. For qualitative imaging like that presented here, this correction may be less important than in a quantitative imaging application; however, this correction should still be performed in order to obtain a more accurate zero signal level and enable construction of signal isosurfaces or contours.

After correcting for phosphor decay and background signals, images are corrected for laser-sheet spatial non-uniformity by normalizing each image by an average of the top 10 rows of the first 5 images, using images in which no flow is observed. We make no attempt to correct for attenuation along the incident laser beam path due to absorption. Lastly, for images presented as stills and in a movie, shot-to-shot variations in laser energy, tracer number density, and temperature are performed by normalizing each image in each field of view by a measure of average LIF signal from a 10 × 10 pixel region near the top of each image. By correcting for shot-to-shot variation in laser energy via a measure of energy from the image itself, variations in signal due to non-uniformities within the 10 × 10 pixel window or major differences in signal within a given image (e.g., an image with a shockwave, as shown in Fig. 3b) will affect the normalization of each image and may result in differences in the relative signal between images within a sequence of images; even so, the scheme adopted in this work is convenient and enables an entire sequence of images to be visualized on the same scale.

Timing of the laser, camera, shutter, and expansion tube operation is summarized in Fig. 2. Time-zero (\(t_{{\rm o}}\)) occurs when the shock wave passes over a shock sensor located near the primary diaphragm station. Before a run, the laser and camera systems are clocked by a SRS delay and pulse generator at 20 kHz. A run is initiated by venting a double-diaphragm section in the tube (see [34] for details of the expansion tube facility and its operation), causing the primary diaphragm to burst and the shockwave to form. The mechanical shutter (the EOPC model SH-20 described above) opens concurrently with the vent; the shutter takes 40 ms to fully open. At this time, the laser pulses illuminate the field of view, but no images are acquired. Roughly 50 ms after the venting of the double diaphragm, a mechanical relay switches the laser and camera clock to a BNC Model 555 delay and pulse generator, which waits to be triggered by the \(t_{{\rm o}}\) signal from the first shock counter; at this time, the laser is not being triggered so no light is emitted from the laser. Once the shockwave arrives at the first shock counter, the BNC delay generator is triggered, clocking the laser and camera at 20 kHz until 360 images are acquired. At this same time (\(t_{{\rm o}}\)), the solenoid valve in the flat plate is also opened, allowing jet fluid to flow through the plate and out into the test section. The test gas arrives in the test section at around t = 6 ms, and images are acquired until 18 ms.
Fig. 2

Schematic of timing of laser, camera, and expansion tube operation

3 Imaging results

In this section, we present a series of images and corresponding pressure and signal-to-noise ratio (SNR) time-histories. A movie displayed at 10 frames per second containing all acquired images can be found in supplementary material.

As mentioned, image acquisition and the flow of jet fluid begin well before the arrival of the shockwave and crossflow (Fig. 2). A sequence of images of the jet start-up process and undisturbed jet are presented in Fig. 3a (and the supplementary movie). The two fields of view are highlighted in the figure by the blue dotted box (upstream) and purple dotted-dashed box. These two fields of view are acquired during subsequent shots in the expansion tube and stitched together in post-processing. The average of all images from \(t=1.5\) to 5.0 ms (75 images) of the undisturbed jet is shown on the far right of Fig. 3a. We observe the starting transient of the jet, the rough location of the barrel shock and Mach disk can be inferred, and downstream unsteady dynamics and instabilities are visualized (e.g., \(t=0.80\) ms in Fig. 3a).
Fig. 3

Image sequences of jet start-up and jet in crossflow. Images are not spaced evenly in time. All presented images are composite of upstream (blue dashed box) and downstream (purple dashed-dotted box) fields of view described in Sect. 2 and Fig. 1. Far right image in both sequences are averages over the indicated time. SNR and average signal values are taken from the 10 × 10 pixel region marked in the left-most image of a. a Instantaneous and time-average (far right) images of starting jet. Red square indicates \(10 \times 10\) pixel region in which average signal level and \(SNR\) is calculated. b Instantaneous and time-average (far right) images of jet in crossflow. At \(t=5.9\) ms, the shock wave can be seen around \(\frac{x}{D}=10\)

At time \(t=5.9\) ms, the shock and subsequent crossflow arrive, as shown in Fig. 3b (and the supplementary movie). Smith and Mungal [2] provide a thorough investigation and description of the behavior of jets in crossflow; for this configuration, the jet-to-freestream momentum flux J, velocity r, and density s ratios are 5.5, 1.5, and 2.5, respectively, at \(t=6\) ms. The Reynolds number \(Re_D,\) using the diameter of the jet (2 mm) and the properties of hydrogen at the throat of the jet is about 800,000.

The upstream half of the first image in Fig. 3b shows the jet immediately after the shock has passed through the field of view, and around \(\frac{x}{D}=10,\) the shock can be seen. To reiterate, the sequence of images in the upstream and downstream fields of view are acquired during different expansion tube shots, and in both shots, the shock was in roughly the same position. The underexpanded jet fluid far from the plate (i.e., near \(\frac{y}{D}>10\)) appears to be translated by the drift velocity of the passing shock, whereas closer to the issuing orifice, the jet has not yet changed trajectory. We also observe the shock is no longer perpendicular to the plate, having been affected by underexpanded jet structure near the plate.

For the rest of the test time, some unsteadiness in the freestream and overall jet trajectory are observed, and typical features of a jet in crossflow are visualized, specifically jet shear-layer vortices on the windward side of the jet and some tornado-like wake vortices in the far field (visible in the supplemental video). The high acquisition rate provides a smooth time-average of the jet (far right of Fig. 3b). A significant difference in the near- and far-field signal intensities are observed, corresponding to the barrel shock and Mach disk (near field) and the mixing of the crossflow with the injectant (far field).

Time-histories of pressure (measured at the location of the last shock counter in Fig. 1), and temperature inferred from pressure (i.e., isentropic expansion after the passing of the shock wave) as well as fluorescence signal \(S_{{\rm f}}\) and spatial \(SNR\) are provided in Fig. 4. Pressure and temperature time-histories are displaced in time to align with the arrival of the shock in the images. Signal is measured in a 10 × 10 pixel region in the image before correction for number density and laser energy are made (the red box in Fig. 3a); the spatial \(SNR\) is computed as the average signal in the same 10 × 10 pixel region divided by the standard deviation of the signal. Signal \(\frac{S_{{\rm f}}}{S_{{\rm f}_{t=0}}}\) is normalized by the measured signal at \(t=0.\)
Fig. 4

Time-histories of (top) static pressure, estimated temperature, (bottom) LIF signal, and spatial \(SNR\)

The LIF signal of toluene is dependent on both temperature \(T\) and toluene number density \(n_{\mathrm{toluene}}\), and in the weak excitation limit is described by
$$\begin{aligned} S_{{\rm f}} \propto n_{\mathrm{toluene}} \sigma (T) \phi (T) \frac{E}{h \frac{c}{\lambda }} \eta , \end{aligned}$$
where \(\sigma\) and \(\phi\) are the absorption cross section and FQY of toluene, \(E\) is the laser fluence, \(h\) is Planck’s constant, \(c\) is the speed of light, \(\lambda\) is the wavelength of excitation light (266 nm), and \(\eta\) is the collection efficiency of the imaging system. For toluene, \(\sigma\) is weakly dependent on temperature, whereas \(\phi\) is an exponential function of temperature, dropping nearly three orders of magnitude from 300 to 900 K [13]. The number density of toluene depends on both pressure and temperature (\(n=\frac{p}{RT}\)). Therefore, the gas dynamic processes that occur in the expansion tube during the course of a run change both the temperature and pressure, which have opposite effect on LIF signal; a reduction in pressure will decrease signal due to decreased number density of toluene, but a reduction in temperature will increase LIF signal (at constant tracer number density) due to an increase in quantum yield. This effect is directly observed in the first image in Fig. 3b, as the temperature increase across the shock has a larger detrimental effect on the LIF signal compared with the increase in number density across the shock. The region behind the shock appears darker relative to the upstream half of the field of view due to the method chosen to correct for variations in laser energy and LIF signal due to changes in tracer number density and temperature.

These changing freestream conditions (i.e., pressure and temperature) enable us to observe changes in the relative LIF signal level and corresponding changes in spatial \(SNR.\) Upon arrival of the shock, signal and \(SNR\) decrease, largely due to an increase in temperature and the high sensitivity of toluene quantum yield to temperature; the decrease in \(SNR\) is evident to the eye in Fig. 3b for \(t=8.4\) ms, as this image is notably more grainy than images acquired at other times. Fluorescence lifetimes are short (<100 ns) for toluene and the excitation laser pulse width is 6 ns, and so degradation in image quality is a result of decreased \(SNR\) and not motion blur. Around 8.5 ms, an expansion fan arrives (having reflected off the end wall of the driver section of the expansion tube), dropping both the temperature and pressure, and signals and \(SNR\) increase again until all the toluene-seeded test gas has passed through the test section (\(t=18\) ms). \(SNR\) up to 30 is observed, and \(SNR\) does scale roughly nearly with \(\sqrt{S_{{\rm f}}},\) suggesting the system is approaching the shot-noise limit, in terms of its spatial noise characteristics.

Lastly, continuous high acquisition rate imaging enables inference of the test gas flow steadiness (in terms of momentum flux) and duration. To do this, we stack up the images and volume render isosurfaces of LIF signal (\(S_{{\rm f}}=0.25\)), and Fig. 5 presents two different views of the jet volume rendered from immediately after the arrival of the shock wave (\(t=6\) ms) until the end of the test time (18 ms). A 5-image moving average is used to smooth the images, and images from every 2 ms (40 images) are colored differently. The first view (Fig. 5a) presents an end-on view of the isosurfaces, which have been rendered translucent. Immediately after the arrival of the shock wave (the red isosurface, extending from \(t=6\) to \(t=8\) ms), the jet is deflected, and from Fig. 5a, we see the jet follows roughly the same trajectory (as defined by the overlap of the translucent surfaces) for nearly the entire time the crossflow is on.
Fig. 5

Volume renderings of the jet in crossflow. a End-on view of isosurface defined by \(S_{{\rm f}}=0.25\) from \(t=6\) ms until the end of the test time (18 ms). Each 2 ms block of data is in a different color, and isosurfaces are translucent. b top-down view of b, showing the downstream extent (\(\frac{x}{d}\)) of the jet as a function of time

The trajectory or penetration of a jet in crossflow is largely dictated by entrainment and therefore \(J\), the momentum flux ratio [2], which for a compressible flow can be written as the ratio of \(P\gamma M^2\) (where \(P\) is pressure, \(\gamma\) is the ratio of specific heats, and \(M\) is Mach number) for the jet and freestream. The stagnation pressure of the jet decreases by roughly 0.5 % per ms, measured by a high-bandwidth piezo-electric transducer in the jet plenum (reduction of 9 % over 18 ms test time), so the momentum flux of the jet decreases 9 % over the 18 ms test time. Because the trajectory of the jet stays roughly the same through the entire time the test gas is flowing, we can infer the freestream test gas has roughly the same momentum flux during this time despite its changing pressure and temperature.

By viewing this isosurface from above, it first appears as though there is some time-dependent structure or dynamic in the downstream extent of the jet plume. However, we cannot comment on this aspect of the jet in this particular study because the upstream and downstream fields of view (split near \(\frac{x}{d}=6\)) are acquired during different tests. Even so, by inspecting the isosurface from the top, we can again see that the trajectory of the jet (as defined by the top portion of the signal isosurface) is roughly constant throughout the test time.

Jet trajectories in previous works have typically been quantified using measures of jet fluid concentration or velocity within the plume of the jet [2]. This work qualitatively marks the jet trajectory using a contour of signal, which may be decreased due to a decrease in temperature or tracer number density, which we cannot distinguish with uncalibrated, single-camera tracer-based PLIF. Therefore, the scientific impact on the fluid mechanics of the jet in crossflow is limited in this particular demonstration. By utilizing a dual-band or multi-camera high-repetition-rate tracer-based PLIF imaging strategy [13, 20, 22], quantitative images of temperature can be acquired, and the ease of use of high-repetition-rate tracer-based PLIF imaging enables these techniques to be used to study flows generated in impulse facilities. The operation of expansion tube facilities at low-enthalpy conditions in standard operating mode or ‘shock-mode’ produces long-duration compressible flow fields, and these relatively long-duration flows provide ample time to establish steady-state conditions for many flows, and high-repetition-rate systems allow meaningful statistics to be collected. Because these facilities allow for custom mixtures of test gases, tracer-based PLIF is a useful technique to study these flows. The combination of high-repetition-rate, quantitative tracer-based imaging techniques with impulse facilities yields opportunities to further study canonical fluid mechanics topics (e.g., jets in supersonic crossflow) or applied configurations (e.g., the geometry of a pylon injector in a scramjet).

4 Conclusion

This work demonstrates the use of turn-key laser and imaging systems to continuously image and visualize an underexpanded jet in crossflow generated in an impulse facility using tracer-based PLIF. A jet of fuel injected into nearly sonic crossflow behind an incident shockwave was imaged using a turn-key 20 kHz, pulsed, 266-nm light source, and an off-the-shelf high speed, CMOS camera coupled to an intensified relay optic. Spatial signal-to-noise ratios between 10 and 30 are observed. The methodology demonstrated within can be extended to quantitative imaging (e.g., thermometry) through the use of dual-camera techniques [13, 20, 22]. High-speed PLIF imaging enables the study and characterization of a variety of practical flows, including high-speed or supersonic flows. Lastly, the ease of use of the equipment used in this work will enable the study of more complex flow fields due to the decreased demand on the experimentalist to manage both the diagnostics and the experimental facilities.

Notes

Acknowledgments

V. A. Miller is supported by the Claudia and William Coleman Foundation Stanford Graduate Fellowship; V. A. Troutman is supported by the Gabilan Stanford Graduate Fellowship. This work is made possible by the Air Force Office of Scientific Research (AFOSR) with Dr. Chiping Li as technical monitor.

Supplementary material

340_2014_5849_MOESM1_ESM.mp4 (4 mb)
Supplementary material 1 (mp4 4131 KB)

References

  1. 1.
    T. Fric, A. Roshko, J. Fluid Mech. 279(1) (1994). doi:10.1017/S0022112094003800
  2. 2.
    S.H. Smith, M.G. Mungal, J. Fluid Mech. 357, 83 (1998). doi:10.1017/S0022112097007891 CrossRefADSGoogle Scholar
  3. 3.
    A.R. Karagozian, Prog. Energy Combus. Sci. 36(5), 531 (2010). doi:10.1016/j.pecs.2010.01.001
  4. 4.
    K. Mahesh, Annu. Rev. Fluid Mech. 45(1), 379 (2013). doi:10.1146/annurev-fluid-120710-101115 MathSciNetCrossRefADSGoogle Scholar
  5. 5.
    C. Nordeen, D. Schwer, F. Schauer, J. Hoke, T. Barber, C. Baki, in 47th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit (AIAA 2011–6045, San Diego, California, 2011), pp. 1–17. doi:10.2514/6.2011-6045
  6. 6.
    G. Kychakoff, R.D. Howe, R.K. Hanson, Appl. Opt. 23(5), 704 (1984)CrossRefADSGoogle Scholar
  7. 7.
    B.K. McMillin, J.M. Seitzman, R.K. Hanson, AIAA J. 32(10) (1994). doi:10.2514/3.12237
  8. 8.
    K. Kohse-Höinghaus, J.B. Jeffries, Applied Combustion Diagnostics (Taylor & Francis, London, 2002)Google Scholar
  9. 9.
    D.A. Rothamer, J.A. Snyder, R.K. Hanson, R.R. Steeper, Appl. Phys. B 99(1–2), 371 (2009). doi:10.1007/s00340-009-3815-2 ADSGoogle Scholar
  10. 10.
    B.J. Kirby, B.K. Hanson, Proc. Combust. Inst. 28(1), 253 (2000). doi:10.1016/S0082-0784(00)80218-0 CrossRefGoogle Scholar
  11. 11.
    M. Tanahashi, S. Murakami, G.M. Choi, Y. Fukuchi, T. Miyauchi, Proc. Combus. Inst. 30(1), 1665 (2005). doi:10.1016/j.proci.2004.08.270
  12. 12.
    C.C. Rasmussen, S.K. Dhanuka, J.F. Driscoll, Proc. Combust. Inst. 31(2), 2505 (2007). doi:10.1016/j.proci.2006.08.007
  13. 13.
    W. Koban, J.D. Koch, R.K. Hanson, C. Schulz, Phys. Chem. Chem. Phys. 6, 2940 (2004). doi:10.1039/b400997e
  14. 14.
    J. Yoo, D. Mitchell, D.F. Davidson, R.K. Hanson, Exp. Fluids 45, 751 (2010). doi:10.1007/s00348-010-0876-2 CrossRefGoogle Scholar
  15. 15.
    J. Yoo, D. Mitchell, D.F. Davidson, R. Hanson, Shock Waves 21(6), 523 (2011). doi:10.1007/s00193-011-0338-7 CrossRefADSGoogle Scholar
  16. 16.
    A. Lozano, S.H. Smith, M.G. Mungal, R.K. Hanson, AIAA J. Tech. Notes 32(1), 218 (1993)CrossRefADSGoogle Scholar
  17. 17.
    N. Jiang, J. Bruzzese, R. Patton, J. Sutton, R. Yentsch, D.V. Gaitonde, W.R. Lempert, J.D. Miller, T.R. Meyer, R. Parker, T. Wadham, M. Holden, P.M. Danehy, Exp. Fluids 53(6), 1637 (2012). doi:10.1007/s00348-012-1381-6 CrossRefGoogle Scholar
  18. 18.
    B. Peterson, V. Sick, Appl. Phys. B 97(4), 887 (2009). doi:10.1007/s00340-009-3620-y CrossRefADSGoogle Scholar
  19. 19.
    R.L. Gordon, C. Heeger, A. Dreizler, Appl. Phys. B 96(4), 745 (2009). doi:10.1007/s00340-009-3637-2
  20. 20.
    M. Cundy, P. Trunk, A. Dreizler, V. Sick, Exp. Fluids 51(5), 1169 (2011)CrossRefGoogle Scholar
  21. 21.
    B. Peterson, E. Baum, B. Böhm, V. Sick, A. Dreizler, Proc. Combus. Inst. 34(2), 3653 (2013). doi:10.1016/j.proci.2012.05.051
  22. 22.
    V.A. Miller, M. Gamba, M.G. Mungal, R.K. Hanson, Exp. Fluids 54(6), 1539 (2013). doi:10.1007/s00348-013-1539-x CrossRefGoogle Scholar
  23. 23.
    M. Gamba, V.A. Miller, M.G. Mungal, in Turbulence Shear Flow Phenomena (Poitiers, France, 2013)Google Scholar
  24. 24.
    B.H. Cheung, R.K. Hanson, Appl. Phys. B 98, 581 (2010)CrossRefADSGoogle Scholar
  25. 25.
    M. Gamba, R.K. Hanson, in 16th International Symposium on Applications of Laser Techniques to Fluid Mechanics (Lisbon, Portugal, 2012)Google Scholar
  26. 26.
    N. Jiang, W.R. Lempert, Opt. Lett. 33(19), 2236 (2008)Google Scholar
  27. 27.
    N. Jiang, W.R. Lempert, G.L. Switzer, T.R. Meyer, J.R. Gord, Appl. Opt. 47(1), 64 (2008)Google Scholar
  28. 28.
    R.L. Trimpi, A Preliminary Theoretical Study of the Expansion Tube, a New Device for Producing High-Enthalpy Short-Duration Hypersonic Gas Flows. Technical Report (NASA, 1962)Google Scholar
  29. 29.
    S. Faust, T. Dreier, C. Schulz, Chem. Phys. 383(1–3), 6 (2011). doi:10.1016/j.chemphys.2011.03.013
  30. 30.
    J.A. Snyder, Development and Application of Tracer-Based Planar Laser-Induced Fluorescence Imaging Diagnostics for HCCI Engines. Ph.D. thesis, (Stanford University, 2010)Google Scholar
  31. 31.
    V.A. Miller, Exp. Fluids 54(7), 1566 (2013). doi:10.1007/s00348-013-1566-7 CrossRefGoogle Scholar
  32. 32.
    E. Kristensson, M. Richter, S. Pettersson, Appl. Opt. 47(21), 3927 (2008)Google Scholar
  33. 33.
    E. Berrocal, E. Kristensson, P. Hottenbach, M. Aldén, G. Grünefeld, Appl. Phys. B 109(4), 683 (2012). doi:10.1007/s00340-012-5237-9 CrossRefADSGoogle Scholar
  34. 34.
    W.N. Heltsley, J.A. Snyder, A.J. Houle, D.F. Davidson, M.G. Mungal, R.K. Hanson, in 42nd AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit (Sacramento, California, 2006), AIAA-2006-4443Google Scholar
  35. 35.
    P. Hoess, K. Fleder, in 24th International Congress on High-Speed Photography and Photonics, Proceedings of SPIE, vol. 4183 (2001), p. 127Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  • V. A. Miller
    • 1
  • V. A. Troutman
    • 1
  • M. G. Mungal
    • 1
    • 2
  • R. K. Hanson
    • 1
  1. 1.High Temperature Gasdynamics LabStanford UniversityStanfordUSA
  2. 2.School of EngineeringSanta Clara UniversitySanta ClaraUSA

Personalised recommendations