Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 136
Twenty-First Symposium on NAVAL HYDRODYNAMICS 24-Inch Water Tunnel Flow Field Measurements During Propeller Crashback C.-W.Jiang,1 R.Dong,2 H.-L.Liu,1 M.-S.Chang1 (1David Taylor Model Basin,2 Johns Hopkins University, USA) ABSTRACT Particle Displacement Velocimetry (PDV) is used for measuring the ring vortex flow field near a propeller tip during crashback. The experiment was conducted in the David Taylor Model Basin (DTMB) 24-inch Water Tunnel with Propeller 4381. The formulation and dispersion of the unsteady ring vortex are presented for two crashback conditions. This test demonstrated that the flow is unsteady even when the propeller is operated in the steady crashback condition. The propeller tip-region-flow velocity and vorticity are determined for several time instances at each different advance ratio. These results can be used to validate CFD predictions. The unsteady ring vortex movements are related to the frequency of propeller transverse forces. Several difficulties with the present system have been identified and recommendations are provided for future improvements. INTRODUCTION A most striking aspect of the propeller flow observed during crashback operation is the formation of an unsteady ring vortex. The existence of ring vortex was observed in the wind tunnel test by Lock  and the complete flow cycle of propeller operation was discussed by Durand . The highly unsteady motion of the ring vortex was observed and visualized by towing tank  and water tunnel experiments at DTMB. Although some discrepancies are noted in measurements obtained from those different experiments, all data show following conclusions: First, the ring vortex is highly unsteady, and yet posseses a dominant frequency of movement and dispersion. This was observed using an air injection technique by which air bubbles were injected from upstream. These bubbles were trapped inside the ring vortex to indicate the location and the existence of the ring vortex as shown in Fig.1 for a generic skew propeller. These pictures were taken from the tunnel side window at 0.2 second intervals to show the continuous movements of the ring vortex during crashback operation. The formation of the ring vortices and its dynamic unsteady behavior are clearly demonstrated. The ring vortex oscillates vertically and horizontally as will be shown more clearly in the later section. Many times the vortex seems to start at the propeller tip and move away, only to return later and complete its cycle of motion. Also, the ring vortex would sometimes suddenly disperse into a mayhem of random cloudy air particles only to reform later into a moving ring vortex again. The trajectory and shape of the ring vortex are highly irregular and yet they are periodical. Second, the periods and the motions of the ring vortex are governed by the reversed mass flow through propeller disk. Third, the propeller exerts a large out-of-plane force rotates around the shaft with the same frequency of the ring vortex motion. Fig.2 is a sample of the measured transverse force from shaft dynamometer during a crashback test. These force time histories are normalized by the mean thrust at the given test condition. It is seen that the force is unsteady with an unsteady peak to peak value of approximately one quarter of the steady thrust for the Propeller 4381. It is also seen that the force is periodical. The period of the force has been examined and it matches that of the moving ring vortex. In the past few years, researchers have been fascinated by those observations and consequently efforts were devoted towards the understanding and modeling of these unsteady phenomena. The flow around a propeller blade undergoes three different physical processes during crashback as shown in Fig.3. (I) In the beginning of braking, a propeller turns in its normal direction and
OCR for page 137
Twenty-First Symposium on NAVAL HYDRODYNAMICS Figure 1. Photos of propeller ring vortex.
OCR for page 138
Twenty-First Symposium on NAVAL HYDRODYNAMICS the flow around the blades is fully attached to the blade surfaces. The flow between the propeller blades moves towards the stern as designed. (II) As the braking procedure continues, the propeller begins to turn opposite to its normal direction and the flow around the propeller blades starts to separate. Also, the flow begins to reverse upstream toward the bow. (III) As the braking procedure continues even further, the flow may reattach to the blade surfaces. At this time, the reversed propeller flow should be well established and extend to a location far ahead of the propeller toward bow. Flow visualization has shown that unsteady ring vortices are present during the second stage of the crashback process; that is, (II) above, while the flow inside the propeller disk was reversed. Figure 2. Time history of propeller unsteady side force. Attempts were made towards the modeling of the unsteady flow field. For example, a two dimensional quasi-steady approach was investigated by the authors. In that approach, a two-dimensional RANS code was exercised and the propeller was replaced by a specified time-dependent body force obtained from propeller lifting surface code. The locations of the computed ring vortex were examined and the side forces were estimated. Fig.4 presents a typical calculated result showing the formation of a vortical structure at J=–0.5. The advance ratio, J, is defined as Vo/nD, where Vo is the ship (or tunnel) speed, n is the propeller revolution per second, and D is the propeller diameter. Although, those results show the observed features of a moving vortex field, it cannot explain the origin of unsteadiness. Detailed quantitative measurements are needed for better understanding the phenomena and for advancing numerical modeling. The objective of the present research program was to measure the propeller induced flow field at the blade tip region. This paper presents the unsteady flow structures visualized by the use of laser sheet as well as the vortex velocity field measured by Particle Displacement Velocimetry. Figure 3. Flow field during propeller crashback. Figure 4. Numerically simulated velocity field at crashback, J=–0.5.
OCR for page 139
Twenty-First Symposium on NAVAL HYDRODYNAMICS Particle Displacement Velocimetry (PDV or PIV), a quantitative flow visualization technique, has been used to measure flow separation and vortex structure in the DTMB towing tank [4,5] and in the DTMB rotating arm facility . Although it is the same technology used to provide flow field information, the flow phenomenon in this case is more unsteady and complex than the previous ones. The present research program thus also serves to explore the capability of PDV in the water tunnel facility. TEST FACILITY AND PDV SETUP The DTMB 24-inch Variable Pressure Water Tunnel was used to conduct this test. This tunnel has a 24-in diameter open-jet test section. This test was conducted without a body and screen-generated wakes. The propeller was driven from the downstream shaft and the rotational direction was reversed to simulate crashback operation. In this test, only steady state braking was studied. Propeller RPM and tunnel velocity were varied to obtain different braking conditions. Propeller 4381 was chosen for this study. Geometric characteristics of this propeller are shown in Table 1. Table 1. Geometric characteristics of Propeller 4381. r/R Pitch/D Chord/D Skew 0.20 1.265 0.1735 0.00 0.30 1.345 0.2283 0.00 0.40 1.358 0.2750 0.00 0.50 1.336 0.3125 0.00 0.60 1.280 0.3375 0.00 0.70 1.210 0.3475 0.00 0.80 1.137 0.3342 0.00 0.90 1.066 0.2808 0.00 0.95 1.033 0.2192 0.00 1.00 1.001 0.0000 0.00 Velocity measurements were performed using PDV with the experimental setup shown in Fig. 5. The light source was an air-cooled, 15W, continuously pulsed copper vapor laser (511 and 578 nm were the primary wavelengths). Timing for laser pulses was synchronized by a PC-based control system. The laser beam was expanded to a 6-mm thick light sheet by a combination of cylindrical and spherical lenses. This light sheet illuminated a desired section of the propeller flow field through a window on the bottom of the 24-inch WT test section. Data was recorded by a 70-mm Hasselblad camera equipped with a 60-mm lens on TMAX 3200 film. This camera was located outside the test section. PDV images were taken at the lower part of the test section where the laser light was not in shadow. The data was recorded with three exposures per image with a delay between exposures at 800 to 960 µsec. The tracer particles used were 40–60 µm in diameter, neutrally buoyant, and fluorescent. The specific gravity of these particles varied between 0.95–1.05. The fluorescent dyes imbedded in these particles responded with green light excitation in the 550–560 nm range of the laser. The particles were seeded upstream from the test model during the image recording period. Figure 5. PDV experimental setup. UNSTEADY FLOW BEHAVIOR As shown in Fig.1, dynamic behavior of the ring vortex is very unsteady. In order to better understand the generation and propagation mechanism of this unsteady ring vortex, a technique was designed to record the flow near the and at the propeller tip at different J values using laser sheet as a light source. A video recorder mounted outside the tunnel recorded the flow while the pressure inside the tunnel was lowered until the flow structures became clearly visible from the trajectories of the cavitation bubbles. Figs.6 and 7 are samples of those recordings. The upstream tunnel is to the left in these pictures. Fig.6 presents the sequence of propeller ring vortex motion in the propeller tip region at J=– 0.472. The flow behaviors in these pictures are
OCR for page 140
Twenty-First Symposium on NAVAL HYDRODYNAMICS Figure 6. Time sequence photos of Propeller 4381 at J=–0.472.
OCR for page 141
Twenty-First Symposium on NAVAL HYDRODYNAMICS Figure 7. Time sequence photos of Propeller 4381 at J=–0.732.
OCR for page 142
Twenty-First Symposium on NAVAL HYDRODYNAMICS typical for this J value. As shown in these pictures, a ring vortex is at the propeller tip at t=0.2 sec. and moves downstream and away from the tip. The ring vortex becomes more and more diffused as it moves away. At t approximately 2.0 sec., the vortex breaks into two due to the strong reversed axial flow near the tip. With further increase of the reversed flow at the tip, t=2.2 and 2.4 seconds, the original vortex starts to move back and the two vortices come close to each other. At t=2.8, one can again observe a single vortex at the propeller tip and the process repeats itself after that. The period of this process from those recordings is estimated at about 2.5 seconds. The period of the unsteady force obtained from spectral analyzed data for this same J is 2.1 seconds. They are in good agreement, in view of the randomness in the flow structures. Fig.7 is the ring vortex motion at J=–0.732. One sees basically the same flow motions as in Fig.6 but quantitatively they are different. First, the period of the vortex motion is less at J=–0.732 and the movement of the vortex is less random; at J=–0.732 the flow is almost identical at t=0.0 and t=1.6 seconds. The spectral analyzed force obtains 1.6 sec. period also. Second, the movement of the vortex is less for J=–0.732 than J=–0.472. For the former (J=– 0.732), it moves only from the 0.6 radius (propeller radius) to 0.9 radius at the flow direction and from 1.5 radius to 2.0 radius vertically. For the latter (J=– 0.472), it moves from the propeller plane to 1.0 radius downstream and from propeller tip to 2.6 radius downward. Third, the shape of the cavitations at the propeller tip indicates the reversed axial flow is stronger at J=–0.472 than that at J=–0.732. These observations could support the conjecture that, for crashback operation, reversed flow is the main governing parameter. In general, the frequency of the propeller unsteady force at crashback operation can be non-dimensionalized by the inflow velocity and propeller diameter. This non-dimensionalized number, ωD/Vo, is about 0.5 at –2<1/J<–1 for the Propeller 4381. PDV IMAGE PROCESS The PDV images collected during the test were digitized and analyzed to obtain quantative flow information including velocity vectors and vorticities. A review of the PDV technology can be found in Adrian 's paper [7,8]. The fundamental principles on which the analysis is based are given in References 4 thru 6. The images taken during the test were digitized by a Leaf Scanner to an array size of 1972×1747 pixels with resolution 250 pixels per inch. Digitized images were enhanced locally, with 64×64 pixels interrogation area to deal with the variation of background non-uniformity. A typical triple exposure image is shown in Fig.8 at J=–0.732 which corresponds to t=0.0 in Fig.7. It is seen that the vortex structure is clearer in PDV image than that in cavitation image. The shaded area at the lower right corner is caused by tunnel blockage in which the light sheet is not transparent. The PDV image velocity was analyzed from the images by an auto-correlation technique that determines the mean displacement of the particles within a given window area. The window area used in this paper is 64×64 pixels and is approximately 2.5×2.5 mm (0.1×0.1 inch) in the real physical space. Also, the distance between adjacent windows was 32 pixels (50% overlap). The window size selected was based on the displacement of the particles. Typically, window size is about three times particle displacement. During image analysis, a velocity vector was assigned to the center of the search area only if a satisfactory correlation pattern was found. Otherwise, the velocity of that area was left as null. The reason that a correlation was not obtained in certain regions could be either a lack of seeding or particle displacement too small to solve the mean displacement. It is very common that the velocities near the vortex center were unresolvable. The uncertainty level of this auto-correlation technique was about 0.4 pixels. For a 20-pixel displacement with 8 pair particles within the window, it resulted in a characteristic error of 2% in velocity. The characteristic error rises in the vicinity of vortex center where displacement is smaller and at the areas where the particle density is lower. A detail uncertainty analysis was described in Dong et al. . The particle tracing method calculates the displacement of single particle trace. The present analysis applied this analysis in the area where auto-correlation technique fails and the particle density is low. It is most useful in the vicinity of vortex center where directions of the velocities vary significantly. VELOCITY AND VORTICITY DISTRIBUTIONS After digitizing and calculating the image data, sample velocity and vorticity distributions are shown in Figs.9 and 10 at J=–0.472 and –0.732, respectively. Contours of the vorticity distribution for the data analyzed are presented at the right of the velocity vector map. The vorticity is normalized by the tunnel speed and propeller diameter. The solid line presents positive vorticity and the dotted line is negative vorticity. The increment between each contour is 10 with –10 and 10 labeled in the figures.
OCR for page 143
Twenty-First Symposium on NAVAL HYDRODYNAMICS The length of the velocity vector is normalized by the inflow velocity which is plotted at the lower right corner of these figures. The velocity vector at every other grid points is presented in the figure for better quantitative presentation. The blade projection is shown as the solid line. Figure 8. Typical flow image at J=–0.732. Fig.9 shows three instances of the analyzed image results at J=–0.472 corresponding to the time near 0.0, 1.8 and 2.6 seconds in Fig.6. Three instances of flow at J=–0.732 are shown in Fig.10 corresponding to the time at 0.0, 0.8 and 1.0 seconds in Fig.7. These results provide instantaneous velocity and vorticity distribution along with the unsteady ring vortex structure. These results show that the flow field can be divided into three different regions: (I) velocity at the same direction as inflow, (II) velocity at the reverse direction as inflow and (III) ring vortex region. PDV analysis shows that the maximum velocities at region (I) are about to twice of the inflow speed for all cases, although their locations vary. The velocity fields in the reversed flow region show large variations and are influenced by the blade loading and the passage of the blade. The influence of blade passage is more clearly seen from the vorticity distributions. Higher vorticities near the blade tip are present in all three time instances for J=–0.472 while they are only weakly present for J=–0.732. This is because the blade is more loaded at J=–0.472. The influences of blade passage in the flow field are seen in J=–0.732. Nevertheless, we can not draw conculsions from those measured results because the laser control device didn't function during the test and consective flow images were not obtained. In the ring vortex region, the vorticities are more uniform than one expected. The vorticity does not seem to vary significantly for different J values and for different locations of the vortex center. The feedback effect of the vortex ring on the blade unsteady loading is more likely from the variation of its location than from the intensity of the vorticity. The flow field measurements collected in the present experiment are limited. It supplies a data base for CFD analysis. Future experiment should be designed to take consective images for detailed flow information. CONCLUSIONS AND RECOMMENDATIONS The 24-inch water tunnel experiment demonstrated that the shedding ring vortex associated with a braking propeller is related to the propeller reverse flow and that the unsteadiness is a speed-dependent phenomena. Typically, a low frequency of about 0.5 Hz is readily observed in the flow field photos and it matches the unsteady frequency of the propeller side forces at corresponding J's. Therefore, the dynamic behavior of the ring vortex is directly related to the unsteady forces that the propeller experiences. The flow field analysis from PDV images can provide quantitative measurements of the velocity and vorticity fields. The results shown in Figs.6 thru 10 contribute not only a data base, but also provide some physical insights for future numerical simulation of this complex flow problem. PDV technology was successfully extended to the DTMB 24-inch water tunnel, and it showed the capability of measuring an unsteady propeller flow field. It is most interesting to see that PDV can capture the blade-tip flow very well, Figs.9 and 10. PDV technique may be used to improve blade-tip design. Valuable lessons were learned in how to physically arrange the necessary hardware, such as lenses, cameras, laser light sheets and seeding injectors. Several areas can be improved for future experiments, e.g., the development of continuous picture and seed injection technique. Synchronized force and flow measurements and automated data recording could also be included. ACKNOWLEDGMENT The work described herein was sponsored by the Office of Naval Research (ONR 334) and performed by the Carderock Division, Naval Surface Warfare Center. The authors would like to thank Mr. David
OCR for page 144
Twenty-First Symposium on NAVAL HYDRODYNAMICS Figure 9. Vector map and vorticity distribution at J=–0.472.
OCR for page 145
Twenty-First Symposium on NAVAL HYDRODYNAMICS Figure 10. Vector map and vorticity distribution at J=–0.732.
OCR for page 146
Twenty-First Symposium on NAVAL HYDRODYNAMICS Walden for his encouragement and support, Dr. Stuart Jessup for his technical discussions and support and James Diggs and Dave Fishpaw for their technical support in carrying out the experiment. REFERENCES 1. Lock, C.N.H., “Photographs of Streamers Illustrating the Flow Around an Airscrew in the ‘Vortex Ring State',” Br. ARC R&M 1167 ( 1928). 2. Durand, W.F. ed., Aerodynamic Theory, Vol. IV, Dover Publication, Inc., New York, ( 1963). 3. Hampton, G.A., “Open Water Force and Moment Characteristics on Three Propellers in a Crashback Condition,” DTMB Report CRDKNSWC/HD-1126– 01 ( 1995). 4. Liu, H.L. and M.S.Chang, “Sailplane Vortex Measured by Particle Displacement Velocimetry,” Symp. on Laser Anemometry, ASME Fluid Engineering Division Summer Meeting, Lake Tahoe, NV ( June 1994). 5. Fu, T.C., “Quantitative Visualization of Three-Dimensional Flow Separation and Vortex Structures by Particle Displacement Velocimetry,” Ph.D. Dissertation, The Johns Hopkins University, Baltimore, Maryland ( 1993). 6. Liu, H.L. and T.C.Fu, “PDV Measurement of Vortical Structures in the DTMB Rotating Arm Facility, ” DTMB Report CRDKNSWC/HD-1416– 02 ( 1994). 7. Adrain, R., “Particle-Imaging Techniques for Experimental Fluid Mechanics,” Annu. Rev. Fluid Mech., Vol. 23, pp.261–304 ( 1991). 8. Landreth, C.C., R.J.Adrain, and C.S.Yao, “Double Pulsed Particle Image Velocimeter With Directional Resolution for Complex Flows,” Experiments in Fluids, Vol. 6, pp.119–128 ( 1988). 9. Dong R., S.Chu, and J.Katz, “Quantitative Visualization of the Flow Structure Within the Volute of a Centrifugal Pump,” J. Fluid Engineering, Vol. 114, pp.390–395 ( 1992).
Representative terms from entire chapter: