Diesel Spray characterization with Schlieren-Mie Technique L. Weiss 1, 2,*, A. Peter 1, 2, M. Wensing 1, 2 1: Dept. of Engineering Thermodynamics, Friedrich-Alexander University Erlangen-Nuremberg FAU, Germany 2: Erlangen Graduate School in Advanced Optical Technologies, Friedrich-Alexander University Erlangen-Nuremberg FAU, Germany * Correspondent author: Lukas.Weiss@fau.de Keywords: Spray, Diesel, Schlieren ABSTRACT Mixture formation determines power, efficiency and emissions in direct injected (DI) Diesel engines. Fuel injection is an important part of this process. Hence, scientists are seeking for advanced measurement techniques to understand the complex spray formation and interaction with ambient gas. The improvement of basic techniques, like Mie and Schlieren, by new measurement equipment is often neglected. High Speed recordings of combined Schlieren and Mie measurements give insights in flow fields and qualitative fuel distribution. A Photron Fastcam SA-Z is used for recordings of 100 000 frames per second at a resolution of 408 x 384 pixels. Powerful light sources provide Schlieren and Mie illumination in two different wavelength (DLR licensed LED-Illuminator 532 nm and Cavilux HF 640 nm). Thus the signals can be separated and analyzed individually. While Mie signals can be easily distracted from the background by setting a proper threshold, a Matlab code was developed to determine the Spray in the Schlieren images. Exemplary the penetration of liquid and non-liquid fuel are plotted and compared to a common spray model by Arai. 1. Introduction Mixture formation determines power, efficiency and emissions in direct injected (DI) Diesel engines. Fuel injection is an important part of this process. Typical injection pressures are between 1200 bar and 2500 bar. Liquid fuel exits the nozzle with up to 500 m/s (Thomas Vogel 2011). The injection event lasts 500 µs to 2000 µs. Up to 9 distinct injections per cylinder cycle are common. The easiest way to characterize the injection is to visualize it. Optical access is achieved through optical engines and spray chambers. In optical engines the pistons and liners are made from glass or sapphire (Carling and Singh 2000). Investigations are very close to the real engine conditions in this case. This is not suitable for basic research, where several environmental conditions want to be set individually. This is why different types of spray chambers were developed. Here fuel is injected into an enclosed environment. The spray propagates without wall or flow interaction, like in an engine.
Independent which type of laboratory is used, liquid fuel is commonly captured with Mie scattered light or by shadowgraph images. Schlieren techniques are applied for non-liquid fuel (Pickett, Kook et al. 2009). Quantitative measurements are not possible since local temperature and species distribution is not known. Both have influence on the signal. Hence, scientists are seeking for advanced measurement techniques to understand the complex spray formation and interaction with ambient gas. Laser induced fluorescence (LIF) and raman spectroscopy (RS) promise quantitative data about temperature and species distribution, but high effort in calibration data is necessary (Egermann, Göttler et al. 2001, Andersson, Hemdal et al. 2006, Linne 2013). Thus these experiments can only be supportive for the basic measurement techniques like Mie, Shadowgraphy or Schlieren. Also the improvement of those basic techniques by new measurement equipment is often neglected. Image quality is extremely high, due to new cameras, light sources and optics. High-speed imaging offers time resolved recording of transient processes. This work shows the setup of a Schlieren experiment to visualize a diesel spray under engine relevant conditions. The Mie signal is superposed and captured with the same camera to achieve a perfect temporal and spatial correlation. Additionally a concept for signal separation is presented. Finally fuel injections of Diesel and gasoline are recorded with a high speed camera with 100.000 fps. 2. Experimental Setup The experimental study was carried out in the High Pressure Combustion Vessel OptiVeP at FAU (see Fig. 1). Fig. 1 High Pressure Combustion Vessel OptiVeP at FAU for basic spray research under diesel engine relevant conditions. (pmax = 100 bar, Tmax = 1000 K, pinj = 4000 bar, experiment frequency = 1 Hz)
This spray chamber is designed for basic spray investigations under diesel engine relevant conditions. Ambient gas conditions can be set individually up to 100 bar and 1000 K. The fuel pressure system provides up to 4000 bar. The used injectors are temperature controlled between - 30 C to 110 C. The chamber is constantly scavenged with air or nitrogen. This offers a 1 Hz repetition rate of experiments. The flushing gas enters the cell in the four top edges, where the electrical heaters are placed. They have a heating power of 32 kw. The gas exits the cell in the four bottom edges where the chillers are placed. The flow speed is about 30 mm/s. The optical access is provided by five quartz windows of 125 mm open diameter. They are 90 mm strong to full fill safety specifications. In this work a special research injector is used. It is provided by Continental Automotive GmbH and is based on a passenger car injector. It has only three holes instead of nine to improve optical access. The injected mass per hole is a representative value for a mid-load operation point of a passenger car diesel engine. Operation Point Tab. 1 Standard operation points at OptiVeP to cover engine relevant conditions. Tgas pgas Tfuel pfuel tinj minj [K] [MPa] [K] [MPa] [µs] [mg] OP 1 873 6 363 120 500 4.1 OP 2 923 6 363 120 500 4.1 OP 3 973 6 363 120 500 4.1 3. Optical Setup The Schlieren setup is based on a z-type arrangement as described at Settles (Settles 2012). Two parabolic mirrors f/# = 10 are used as field elements. Additional planar folding mirrors are necessary to keep the test rig off the optical axes before and after the measuring volume. As Schlieren light source the DLR licensed LED Illuminator is used in continuous mode emitting green light around 532 nm. The light is collimated with a f/# = 1 aspheric lens and focused on a pinhole. Here a virtual point light source is created with defined dimensions. The virtual light source is in focal distance to the field element. After the second field element and second planar folding mirror the light is focused onto a second pinhole, which is used as Schlieren cutoff. A camera objective (Zeiss Makro Planar f=100 mm) projects the signal onto the camera chip. For single shot recordings a PCO SensiCam was used. The focus of the camera lens is set to infinity.
The sharp focus plane of the Schlieren setup is in the middle of the two field elements. Fig. 2 compares images of a symmetrically aligned and two miss aligned cases on the example of a gasoline direct injection. Only in one case Schlieren are visible. The other two images show basically shadowgraph recordings. To superpose the Mie Signal three flash-lamps are mounted normal to the Schlieren light axis. To get sharp Schlieren and Mie pictures the object has to be placed symmetrically between the two field elements in the distance of the focal length of the second field element. Resulting images are shown in Fig. 3, where a diesel spray is injected into nitrogen atmosphere at 923 K and 60 bar. The background Schlieren are much stronger than in Fig. 2 at room conditions, due to local temperature gradients. The fine structured spray differs from the background. The combined image shows the advantage of this technique. Both, liquid and gaseous fuel are visible at the same time from the same point of view. The superposed images are nice to look at, but it is very hard to separate the two signals. Thus the Mie light sources where substituted by a red laser (Cavilux HF; 640 nm). Fig. 2 Images 1-3 show the influence on the schlieren object distance relative to the field elements. A gasoline direct injection at room conditions is visualized in this case.
Fig. 3 Example of a Schlieren Mie superposing image in a Diesel injection at 60 bars and 923 K nitrogen atmosphere (OP2) Then the Schlieren and Mie signals are separated by a dichroic mirror. Two additional mirrors are used to equalize the optical path length. Schlieren cut off is placed in the Schlieren light path only. A stereoscope collects both signals and projects them on the camera chip next to each other. This final setup is shown in Fig. 4. A resulting image is shown in Fig. 5. Again a gasoline spray is shown under room conditions. It has to be noted, that only the upper spray plum is illuminated by the laser. This image also shows, that powerfull lightsources are neccesary. The resulting image can be devided into two images, which can be analyzed individually. Signal seperation improves the image quality, since the Schlieren cut-off does not decrease the Mie signal intensity anymore. It also has to be noted, that the chip resolution has to be high enough to achieve reasonable image resolution. Fig. 4 Based on the basic Setup this setup is designed for High-Speed recordings and signal seperation of combined Schlieren and Mie measuremnts.
18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics LISBON PORTUGAL JULY 4 7, 2016 Fig. 5 Raw image of a gasoline spray injected into room atmosphere. Mie and Schlieren Signal are seperated on the camera chip as described in Fig. 4. Finally a high speed camera is used to record a whole injection. Photons Fastcam SA-Z M1 has a resolution of 1024 x 1024 pixels at 20 000 frames per second (fps). To visualize diesel injection, where fuel exits the nozlle with 500 m/s, even higher frame rates are neccecary. Therefore the camear resolution can be decreased and higher frame rates are possible. Thus recordings at 100 000 fps and 408 x 384 pixels where done. The Schlieren light source was used in continious mode. The shutter time of the camera was set to 1 µs. Exposure of the Mie signal was determined by the puls width of the Laser, which was set to 100 ns. The laser pulses where triggered by the camera. Fig. 6 Single images of high speed recordings show slightly lower signal intensity, than in Fig. 3. Fig. 6 shows single images of high speed recordings. Left and right are the single Schlieren respectively Mie Signals. The center one shows the superpsosed signals. The left plum is the
investigated one. Schlieren pictures show background Schlieren, which are unevidently at 60 bar and 923 K nitrogen atmosphere. A self developed Matlab code extracts the Schlieren Signal from the background. This is not possible by simple threshold separtation. In case of high speed recordings background distraction is possible, since the background schlieren do not move significantly during the injection event. Neverthelss the developed code is also able to seperate the Spray signal from a single image. Based on the processed pictures like in Fig. 7 further analysis are done. Fig. 7 Raw image and cropped Schlieren Signal. The signal determination is done by a self developes Matlab code. 4. Results High-speed recordings in Fig. 8 catch fast movement of structures along the spray axis until they reach the spray tip. Tip fuel is shearing off to the surface zone, after losing all kinetic energy. Succeeding fuel overtakes the spray tip again and again, pushing the spray front further. Sheared fuel stays at the same position until the injection is finished. The red circled structure in Fig. 8 (a) and (b) marks such a phenomena.
Fig. 8 Two succeeding images of a single injection of Diesel fuel into 60 bar and 923 K nitrogen atmosphere. Arrows indicate the flow field of the injection. Red circled areas show pinned sheared off fuel. Fig. 9 compares local appearance possibility of Mie and Schlieren Signal at OP1. While penetration depths accord with each other, cone angles are much wider for Schlieren, even at 150 µs. The mentioned sheared off gas phase leads to higher cone angles, which stay constant over injection. After end of injection increase of the cone angle close to the nozzle is noticeable. No maximum penetration depth is reached. Instead the visualized non-liquid spray front is penetrating further than the fluid phase. Sheared off structures at the spray flank stay constant over time (see 700 µs and 1100 µs). Fig. 9 Comparison of Mie and Schlieren Signal at OP1 in local development over time.
Mean penetration depth of Schlieren and Mie signal is plotted over time, representing non-liquid and liquid phase propagation, see Fig. 10. Both signals follow the same curve until 200 µs after visual start of injection (vsoi). Then fuel penetration reaches equilibrium conditions. Measured spray tip penetration fits to the predictive spray model for liquid fuel penetration from Hiroyasu & Arai (Hiroyasu and Arai 1990), which is one of the best suiting models for that kind of injection (Thomas Vogel 2011). Fig. 10 Temporal evolution of penetration depth of liquid (Mie-Signal) and non-liquid fuel (Schlieren-Signal) compared to the predictive Arai spray model (Hiroyasu and Arai 1990). 5. Conclusion The Schlieren Mie combination offers in situ local detection of liquid and non-liquid phase. With new cameras and light sources single injections can be recorded with up to 100 000 fps. Sheared off non-liquid fuel increases cone angels in the Schlieren Signal. While liquid penetration depth reaches a maximum value due to an equilibrium between evaporating and succeeding fuel, the non-liquid penetration follows the spray model from Hiroyasu & Arai. 6. References Andersson, M., et al. (2006). "Temperature determination based on spectral shift of exciplex fluorescence." ICLASS, Kyoto, Japan(06-165).
Carling, R. W. and G. Singh (2000). Review of heavy-duty engine combustion research at sandia national laboratories, Sandia National Laboratories, Albuquerque, NM (US). Egermann, J., et al. (2001). "Application of spontaneous Raman scattering for studying the diesel mixture formation process under near-wall conditions." SAE Technical Papers. Hiroyasu, H. and M. Arai (1990). "Structures of fuel sprays in diesel engines." SAE Technical Papers: 900475. Linne, M. (2013). "Imaging in the optically dense regions of a spray: A review of developing techniques." Progress in Energy and Combustion Science 39(5): 403-440. Pickett, L. M., et al. (2009). "Visualization of diesel spray penetration, cool-flame, ignition, hightemperature combustion, and soot formation using high-speed imaging." SAE International Journal of Engines 2(1): 439-459. Settles, G. S. (2012). Schlieren and shadowgraph techniques: visualizing phenomena in transparent media, Springer Science & Business Media. Thomas Vogel (2011). Comparison of current Spray models under high pressure and high temperature engine relevant conditions.