1,721,047 research outputs found
An experimental and numerical study on cavitating shear flows.
No full textThe dynamics of incipient and developed vortex cavitation was examined. The present work is divided into two parts: the first part deals with cavitation inception, that was studied numerically, and the second part deals with developed cavitation, that was studied experimentally. Incipient cavitation was studied by using a Direct Numerical Simulations (DNS) method for multi-phase flows developed by Unverdi and Tryggvason (1992). The interaction between non-cavitating and cavitating bubbles and the vortex flow was simulated. Incipient cavitation is difficult to observe experimentally because the minimum length scales of the cavitating flow are very small (the size of cavitation nuclei are often on the order of 10 mum) and because the nuclei capture process is stochastic and occurs at different locations in a turbulent flow. The numerical study analyzed the capture time of the bubbles, the bubble trajectory as it gets entrained by the vortex, the effect of the bubble and vortex core diameter ratio, the effect of the cavitation number, and bubble Weber number. The DNS results were compared with a simple one-way coupled particle-tracking (PT) model, using the equation of Johnson and Hsieh (1966) and the Raleigh-Plesset equation (Plesset, 1948). In the PT model different drag and lift coefficients were used in order to determine which model agrees well with the DNS results. It was determined that the drag coefficient given by Haberman and Morton (1953) gave a trajectory that was closer to the trajectory computed with DNS than the drag coefficient by Clift et al. (1978). A lift coefficient that is between the values given by Saffman (1965) and Dandy and Dwyer (1990) would give better comparison with the DNS results. Developed cavitation in a shear layer was studied experimentally in order to determine the effect that the growth and collapse of cavitation have on the dynamics of shear flows. Planar Particle Imaging Velocimetry (PIV) was used to measure the velocity field, the vorticity, strain rates and Reynolds stresses of the flow downstream of the cavitating and non-cavitating shear layer, and the flow pressures and void fraction were also measured. Cavitating shear flows with different cavitation numbers were compared to the non-cavitating shear flow. It was determined that the cavitating flows have higher levels of vorticity, strain rate, and Reynolds stresses in the core of the shear layer, but the differences were not significant. The average flow field was not significantly altered both in terms of average velocities and mean pressure drop across the test section. Downstream of the shear layer, the development of cavitation reduced somewhat the growth of the shear layer (by a factor of about 5%). The reduction of the shear layer thickness is consistent with the observations of Brown and Roshko (1974), Hermanson and Dimotakis (1989), and Belahadji et al. (1995).PhDApplied SciencesMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/132806/2/9990911.pd
Electrical-impedance tomography for the quantitative measurement of solids distributions in gas -solid riser flows.
No full textAn electrical-impedance tomography (EIT) system was constructed to non-invasively measure distributions of solids particles in the riser of a gas-solid circulating fluidized bed (CFB), intended for the development and validation of computational models of multiphase flows. EIT systems have often been applied to gas-solid flows, mainly yielding qualitative data, but here an EIT system is validated against a reliable gamma-densitometry tomography (GDT) system, yielding quantitative data. The systems were applied to a pilot scale CFB with a 14 cm inner-diameter, 5.77 m high riser circulating fluid catalytic cracking catalyst with air. Overall solids volume fractions were generally less than 25 percent, and the flows examined were annular with a dilute core. Radially symmetric solids volume fraction profiles in the form of fourth-order polynomials were produced by EIT and GDT and mostly agree within uncertainty. The EIT system is considered validated against GDT and can therefore be used in place of the slower, more expensive system. The EIT system can measure a solids distribution in approximately one second with low error, a significant improvement over the GDT system which requires several minutes. EIT overpredicts the solids concentration near the riser walls in comparison to GDT because of a bias error but this is within uncertainty. The bias error is +6% at the riser walls and -4% at the riser center. Both EIT and GDT underpredict the solid loading at the center of the riser, in many cases predicting zero loading where the literature indicates solids volume fractions of a few percent which is within the systems' uncertainty. An invasive method may be needed to measure the solids loading in this region. Mixture models used to relate solids volume fraction to electrical permittivity were examined during the design of the EIT system and assessed for their suitability. The Rayleigh mixture model was found to be suitable for use with the EIT system. It was also found that the permittivity of relatively dense distributions of particles (solids volume fraction greater than 40 percent) is sensitive to the particle arrangement, which has implications for the application of EIT to dense gas-solid flows.PhDApplied SciencesChemical engineeringMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/124585/2/3150112.pd
Dynamics and noise emission of vortex cavitation bubbles.
No full textThe dynamics and noise emission of cavitation bubbles forming within the core of a single line vortex were examined experimentally and numerically. For the experiment, a steady line vortex was formed downstream of a hydrofoil mounted in the test section of a re-circulating water tunnel, allowing for the detailed examination of the growth, splitting and collapse of individual cavitation bubbles as they experience a reduction and recovery of the local static pressure. The average Reynolds number of the vortices was ReGamma = 250000, and the average core radius was 4.5 mm. The growth and collapse of bubbles with maximum aspect ratios of 60 were examined. The acoustic emissions from the bubbles were detected during growth, splitting and collapse. The impulse produced upon collapse was the strongest and usually 10 times the rarely detected impulse caused by inception and splitting. However, the impulse produced by the elongated bubbles was 100 times less than the impulse produced by near-spherical cavitation bubbles. The equilibrium radius of elongated vortex cavitation bubbles are related to the original vortex properties and the local static pressure. The observed bubble radii were between 10% and 15% of the original vortex radius. The analytically calculated bubble radii are nearly 3 times larger, even after considering the influence of the cavitation bubble on the original vortex flow field. Similar results were found after direct simulation of the axisymmetric bubble dynamics with DF_UNCLE(c). These differences are ascribed to the three-dimensional and wavy interfaces of the experimentally observed bubbles compared to the assumption of axisymmetry used in the calculations. The growth and collapse of individual axisymmetric cavitation bubbles were performed to study scaling effects, and the process of bubble splitting. Finally, cavitation inception related to the interaction of two counter-rotating vortices was experimentally examined. In traditional vortex cavitation scaling, the strongest (primary) vortex is expected to cavitate first. However, vortex interactions can lead the stretching of a weaker (secondary) vortex as a result of Crow-like instabilities. The weaker vortex can then have a substantially reduced core pressure and incept before the primary vortex.PhDApplied SciencesMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/125635/2/3208440.pd
Numerical studies of drop motion in axisymmetric geometry.
No full textThe objective of the present work has been to investigate two major aspects of spray processes: drop collision with a boundary and secondary breakup of drops. A numerical technique based on a finite difference/front tracking method is developed for axisymmetric geometry and the governing equations are solved for both the drops and the surrounding fluid. Inertia, viscous effects and surface tension are fully accounted for. The heat loss of drops impacting on a rigid, cold substrate is examined first. For low Weber numbers, the drop deformation is small and heat losses depend strongly on the Peclet number. For high Weber numbers the drops undergo large deformations and most of the heat is generally lost, with a rate that decreases with increasing Peclet number. The effects of the Reynolds number and wall boundary conditions are also examined. A simple model based on the conservation of kinetic and surface energy is constructed. The model results are in reasonably good agreement with the full simulations. The secondary breakup of liquid drops is then examined for both small (1.15) and large (10) density ratios. The drop is accelerated by either constant acceleration (gradual disturbances) or impulsive acceleration (shock disturbances). The evolution of drops is mainly controlled by the Eotvos number or the Weber number. Depending on these numbers, different breakup modes are observed and the physical mechanism causing the transitions between them is studied. Comparisons with an inviscid model indicate that a backward-facing bag is a viscous phenomenon. The viscosity of fluids, represented by the Ohnesorge number (Oh), plays a secondary role. As Oh increases, the boundaries between different breakup modes move to higher Eo (or We). The effect of the viscosity ratio is noticeable only when it is large. Simulations using the Boussinesq approximation show that results for a small density ratio can be applied to a relatively broad range of small density ratios. At larger density ratios, some differences are seen. Breakup regime maps are presented to summarize the results.PhDApplied SciencesMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/130981/2/9825239.pd
An experimental investigation into the dynamics of propeller tip vortices and the associated cavitation noise.
No full textAn experimental investigation into the tip leakage vortex (TLV) flow from a 3-bladed ducted marine propulsor was performed to assist in understanding the dynamics of 'limited event-rate cavitation' (LEC) inception in the vortex core. The study involved particle image velocimetry (PIV) measurements of the liquid phase TLV synchronized with the propeller angular position. PIV measurements of the tip vortex from the propeller without the duct were also acquired. Additionally, the dynamics of cavitation nuclei capture by a concentrated vortex were investigated experimentally through observations of laser induced bubbles placed near the core of a simple line vortex, and numerically using a particle tracking model for a similar flow. The acoustic noise from cavitation bubbles in a simple line vortex was also studied experimentally. Velocity-inferred vortex core pressures performed here and combined with cavitation inception observations by Chesnakas & Jessup (ASME-FEDSM 2003) indicated that LEC occurred at a location downstream from the blade far from the average location of minimum vortex pressure, and at a higher cavitation inception number than expected. Detailed analysis of the instantaneous PIV fields revealed that the TLV was comprised of a system of multiple vortices. An identification procedure was developed to measure the multiple vortices. Elevated velocity fluctuations coincided with the location along the vortex core where LEC occurred. This suggests that LEC is associated with fine vorticity filaments that are possibly being stretched and undergoing transient pressure dips sufficient for cavitation. This finding contradicts the classical scaling law of McCormick (J. Basic Eng. 1962). The noise produced by collapsing cavitation bubbles inserted optically into the core of a line vortex was examined experimentally, and the produced acoustic pulses upon collapse were found to be broadband (>100 kHz). The noise due to bubble growth is negligible compared to the collapse pulse. The efficiency of vortex cavitation bubbles at transforming their stored mechanical energy into radiated acoustical energy decreased with decreasing cavitation number.PhDApplied SciencesMechanical engineeringOcean engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/123928/2/3106141.pd
Experimental and numerical examination of cavitating flows.
No full textHydrodynamic cavitation occurs when, as a result of high velocity, the local liquid pressure falls below the liquid vapor pressure resulting in regions of vapor. The modeling and simulation of complex cavitating flows requires advances in two main areas. First, the physical nature of these flows must be examined, and the basic mechanisms must be understood. Second, models must be developed which are sufficiently detailed to incorporate the physical processes identified. These tasks are difficult because cavitating flows are unsteady, involve two phases, and occur over a wide range of length and time scales. In the present work, research is presented which is directed toward revealing the basic mechanisms of complex cavitating flows. The work is divided into two sections. In the first section, an experimental examination of a complex cavitating flow is presented. The flow near an attached cavity is captured using Particle Imaging Velocimetry (PIV) and Holographic Particle Imaging (HPI). The formation and structure of the closure region of partial cavities are examined. The bubbly flow downstream of the attached cavity is measured, and the effects of gas diffusion are investigated. Existing models of gas diffusion are compared with the experimentally determined gas diffusion rates. It is determined that models which incorporate the effects of turbulent mixing near the cavity interface over-predict the gas diffusion rate by approximately one order of magnitude. In the second section, three-dimensional, direct numerical simulation of the dynamics of cavitation bubbles are presented. The full Navier-Stokes equations are solved by a finite difference/front tracking method that allows a fully deformable bubble surface. The effects of fluid viscosity, surface tension, and liquid shear are examined for the case of single cavitation bubbles. It is shown that the collapse of a single bubbles is slowed as viscosity is increased. The collapse of a bubble is accelerated when subjected to a shear flow, and bubbles collapsing near a solid boundary may not produce a re-entrant jet when surrounded by a shearing flow. Preliminary two-dimensional multiple bubble simulations are also shown.PhDApplied SciencesMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/129744/2/9610276.pd
LDV measurements and analysis of gas and particulate phase velocity profiles in a vertical jet plume in a 2D bubbling fluidized bed.
No full textFluidized beds are commonly used as chemical reactors and solid fuel combustors, where high-speed gas jets are employed to introduce reactants to the system. Quantitative gas and particulate phase velocity measurements are needed to characterize the transport phenomena of these jet plumes into the particulate emulsion. Two component Laser Doppler velocimetry (LDV) was used to investigate the gas and particulate phase velocities and resulting transport of a vertically injected gas jet plume in a two-dimensional bubbling bed. LDV measurements of this optically dense multiphase flow are challenging due to laser intensity fluctuations, which mix with the recorded burst frequencies. This problem was resolved by optimizing the Bragg Cell configuration and burst signal processing. The jet gas was seeded with ice crystals, and bursts from the bed particles and gas tracers were simultaneously acquired. These bursts were differentiated based on their intensity and coincidence to determine the gas and particulate phase velocities. The behavior of the single-phase gas jet in the empty bed was examined. The self-similar velocity profile growth was consistent the development of a free two-dimensional turbulent jet. The bed was filled with high-density polyethylene microspheres. The gas and particulate phase velocity profiles of the jet were measured. Similarity profiles are presented and the resulting void fraction, mass flow and momentum transport calculations are analyzed. The effect of fluidization velocity on the jet dynamics was examined and was shown to influence the rate of mass entrainment into the jet plume and the momentum exchange between the phases. The ratio of the maximum gas and particulate phase velocities appeared to follow a similar trend for all test conditions. Hence, the ratio of the drag force coupling to the particulate or gas phase momentum is constant at a given axial location from the jet inlet. Additionally, the mass-averaged velocity ratio for the two phases, which is equivalent to the ratio of the centerline velocities, develops in a similar fashion. This phenomenon is observed for high-speed gas jets in bubbling beds with spherical particle entrainment when Rep > 1,000 so that CD ∼ 0.4.PhDApplied SciencesChemical engineeringMechanical engineeringPetroleum engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/127159/2/3429383.pd
Conceptualization, development, and verification of a bubble void fraction instrument.
No full textA December 2002 experimental investigation (coined HIPLATE) was initiated by the University of Michigan on ONR/DARPA funding. The HIPLATE program investigated the physics of drag reduction by microbubble injection at high Reynolds Number and large scale. An Impedance Void Fraction Meter (IVFM) was developed to surmount the shortcomings of the optical techniques. IVFM detects impedance changes and used Maxwell's equation to relate to void fraction by measuring the voltage difference across two electrodes mounted flush to the solid/fluid interface. The IVFM impedance of the bubble-water mixture along with the optical measurements during the HIPLATE help determine void fraction and interface locations. An uncertainty analysis was performed with known resistors. It is observed that total error decreased almost linearly across the low resistance range until around 15 kO and increased mildly above 15 kO. Simulation model of the IVFM using Maxwell's equation could predict the void fraction of water-glass sphere mixture with about 2.4% error, where the diameter of sphere was distributed from 10 to 20% of the electrode size. Simulation model could predict the impedance of the IVFM placed in the prescribed water layer thickness condition within nearly the uncertainty ranges of the experimental setup. A small scale bubble injection test was performed to test the applicability of simulations to dynamic situations. Estimated impedance by simulation using area ratio was than those by experiment about 12% on average. Using the void fraction with the bubbles assumed perfect spheres, the simulation predicted 28.6% smaller impedance change than did the experiments on average. However, the uncertainty intervals of the experimental impedance changes were so large (about 50%) that the applicability of the simulation to the dynamic situation could not be determined from this experiment. Finally, different configurations of the IVFM with multiple pairs of electrodes are shown to have the ability to improve void fraction measurement in the future. The relative motion of the friction and separator plates in wet clutches during the disengaged mode causes viscous shear stresses in the fluid passing through the small gap. This results in a drag torque on both the disks that wastes energy and decreases fuel economy. To understand the effect of number of grooves and groove depth on the drag torque reduction, five different plates were manufactured for experiments with various parameters. The mathematical model enabled us to calculate the drag torque on the disks and an axisymmetry solver verified the solution. 3D models of one grooved and one flat disk are simulated using CFD and the drag torque results matched well with experiment within 20%. Flow visualization images showed when the fully flooded plate started to aerate and how it evolved.PhDApplied SciencesMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/124382/2/3138132.pd
Examination of the effects of subatmospheric pressure on erythrocytes and the formation of cavitation in blood.
No full textIt is currently thought that subatmospheric pressure causes hemolysis, resulting in constraints on the design of blood-handling devices and extracorporeal circulation techniques. The work presented in this thesis stems from the hypothesis that subatmospheric pressure is not a hemolytic mechanism, and thus, can be utilized safely. The first portion of the thesis concentrates on the effect of subatmospheric pressure on erythrocytes. A static pressure apparatus is described that prevents bubble formation in blood at pressures below vapor pressure. A second apparatus is presented for the study of erythrocyte damage by pressure induced pipe flow. The plasma free hemoglobin (PFHB) generated by positive pressure induced flow is compared to that generated by subatmospheric pressure induced flow. The results from these experiments show that subatmospheric pressure is not a hemolytic mechanism. The second portion of the thesis investigates cavitation in blood. To study cavitation inception in blood, a Cavitation Susceptibility Meter (CSM) and a syringe-pump flow apparatus were designed. The CSM and flow apparatus were used in a sheep model to measure the nuclei content of in vivo blood. At tensions 120 kPa, the nuclei content of arterial blood was determined to be 2.7 per liter of plasma, with venous blood having 0.5 per liter of plasma. From bubble stability theory, the nucleus radii were calculated to be 0.3 m. It was also shown that a tension of 120 kPa (127 kPa, gauge) does not cause erythrocyte lysis, further validating that subatmospheric pressure is not hemolytic. To study the hemolytic potential of cavitation, an analytical model was formulated based on the Rayleigh-Plesset equation, and the assumption that erythrocyte lysis is caused by the straining flow generated by a cavitating bubble. To validate the model, in vitro experiments with the CSM and flow apparatus were performed. At cavitation event rates up to 9 events/sec (270 nuclei per liter of plasma), the increase in PFHb was only 1 mg/dl, as predicted by the analytical model. Higher event rates could not be achieved without flow separation. Recirculating studies with the CSM showed that an increase in PFHb correlated with the formation of attached cavitation.PhDApplied SciencesBiological SciencesBiomedical engineeringBiophysicsHealth and Environmental SciencesMedicineUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/131403/2/9909868.pd
Acoustic characterization of cavitation in reverberant environments.
No full textAcoustic localization techniques are applied in a water tunnel to study low event rate cavitation in a vortical flow. The timing and waveform of these cavitation pulses are not known. An array of 16 receiving hydrophones is used to measure acoustic pulses, and array-signal-processing techniques are utilized to estimate the source location in the water tunnel test section. The measured bandwidth of the acoustic pulse from the growth/collapse of a small isolated cavitation bubble is more than 200kHz, and the measured pulse duration is ∼15-20 micro-seconds. The direct path signal between the cavitation source and the receiving hydrophones used for this effort may be obscured by background hydrodynamic noise in the water tunnel. Fortunately some of the direct-path signal arrivals are distinct enough for acoustic detection and localization. The direct-path signals are windowed and cross-correlated to obtain arrival time differences between hydrophones. These arrival time differences are used in conjunction with a simple ray-based acoustic model to estimate the source location in three dimensions. The source location estimate can be used in conjunction with a back-propagation routine, which was developed but not validated fully, to recover part of the original cavitation pulse waveform and improve the localization estimate. Once developed, the acoustic localization method was used in an experimental examination of the locations of cavitation inception and bubble growth, and of the dynamics of cavitation bubbles in a pair of parallel counter-rotating vortices. The underlying vortical flow, static pressure, and nuclei distribution were characterized in separate studies. These fluid and flow parameters influenced the location, acoustic signal, and dynamics of the cavitation bubbles. Details of the acoustic signature of the cavitation bubble were investigated during its inception, growth, splitting, and collapse. In the chosen flow field, it was found that during bubble growth the acoustic signal is the strongest with the bulk of the signal energy in frequencies between 1 and 6kHz. Here the frequency content of the acoustic signal during inception and growth was related to the volumetric change of the bubble measured from images taken with a high-speed video camera with a correlation of 84%.PhDApplied SciencesMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/126640/2/3276109.pd
- …
