David R. Stinebring

A preconditioned Navier–Stokes method for two-phase flows with application to cavitation prediction

Abstract An implicit algorithm for the computation of viscous two-phase flows is presented in this paper. The baseline differential equation system is the multi-phase Navier–Stokes equations, comprised of the mixture volume, mixture momentum and constituent volume fraction equations. Though further generalization is straightforward, a three-species formulation is pursued here, which separately accounts for the liquid and vapor (which exchange mass) as well as a non-condensable gas field. The implicit method developed here employs a dual-time, preconditioned, three-dimensional algorithm, with multi-block and parallel execution capabilities. Time-derivative preconditioning is employed to ensure well-conditioned eigenvalues, which is important for the computational efficiency of the method. Special care is taken to ensure that the resulting eigensystem is independent of the density ratio and the local volume fraction, which renders the scheme well-suited to high density ratio, phase-separated two-fluid flows characteristic of many cavitating and boiling systems. To demonstrate the capabilities of the scheme, several two- and three-dimensional examples are presented.

High Reynolds Number, Unsteady, Multiphase CFD Modeling of Cavitating Flows

A preconditioned, homogeneous, multiphase, Reynolds Averaged Navier-Stokes model with mass transfer is presented. The model is preconditioned in order to obtain good convergence and accuracy regardless of phasic density ratio or flow velocity. Engineering relevant validative unsteady two and three-dimensional results are given. A demonstrative three-dimensional, three-field (liquid, vapor, noncondensable gas) transient is also presented. In modeling axisymmetric cavitators at zero angle-of-attack with 3-D unsteady RANS, significant asymmetric flow features are obtained

Relative blood damage in the three phases of a prosthetic heart valve flow cycle.

Blood flow through a prosthetic heart valve operating in a ventricular assist device can be subdivided into three phases: a) forward flow through an open valve, b) rapid valve closure, and c) regurgitant back flow through a closed valve. Recent studies of fluid stresses in the Penn State Electric Left Ventricular Assist Device (PS LVAD) operating under physiologic conditions indicate that Reynolds stresses of possibly hemolytic magnitude may exist in the valve area. Although several studies have been made of the fluid stresses seen in forward flow through an open valve, few have looked at valve closure or backflow, and none have related these stresses directly to blood damage. In this study, novel in vitro blood flow loops were developed to allow for the separate analysis of the three flow phases of a Bjork-Shiley monostrut Delrin disk valve operating in a PS LVAD. Forward flow through fully open aortic and mitral valves and backflow through closed valves are studied separately in flow loops driven by a roller pump with the LVAD acting as a valve housing and compliance vessel. Valve closure is investigated with a PS LVAD operating in a low volume mock circulatory loop characterized by cavitation potential through stroboscopic videography of this mock loop, using saline as the working fluid. Rate of hemolysis, characterized by the index of hemolysis, IH, is determined for each of the three flow loops charged with fresh porcine blood.(ABSTRACT TRUNCATED AT 250 WORDS)

Two Aspects of Cavitation Damage in the Incubation Zone: Scaling by Energy Considerations and Leading Edge Damage

This study focused on two aspects of the cavitation damage problem, namely an energy approach to the scaling of cavitation damage in the incubation zone and damage near the leading edge of a test model. The damage to the surface of the models was in the form of small indentations in which no material was removed. For a wide range of velocities namely 14.9 to 59.3 m/s the rate of pit formation per unit area in the maximum damage zone increased by the sixth power of velocity. Furthermore it is shown that the damage rate versus velocity data are in good agreement with three other investigations. The volumes of the pits were found to increase by the fifth power of velocity. A relationship between the volume of a pit and the cavitation bubble collapse energy absorbed was developed. The damage to the leading edge was felt to be due to the reentrant jet striking the leading edge of the cavity creating a short term pressure rise causing the collapse of any cavitation bubbles in this area.

A Method for Real-Time In Vitro Observation of Cavitation on Prosthetic Heart Valves

A method for real-time in vitro observation of cavitation on a prosthetic heart valve has been developed. Cavitation of four blood analog fluids (distilled water, aqueous glycerin, aqueous polyacrylamide, and aqueous xanthan gum) has been documented for a Medtronic/Hall prosthetic heart valve. This method employed a Penn State Electrical Ventricular Assist Device in a mock circulatory loop that was operated in a partial filling mode associated with reduced atrial filling pressure. The observations were made on a valve that was located in the mitral position, with the cavitation occurring on the inlet side after valve closure on every cycle. Stroboscopic videography was used to document the cavity life cycle. Bubble cavitation was observed on the valve occluder face. Vortex cavitation was observed at two locations in the vicinity of the valve occluder and housing. For each fluid, cavity growth and collapse occurred in less than one millisecond, which provides strong evidence that the cavitation is vaporous rather than gaseous. The cavity duration time was found to decrease with increasing atrial pressure at constant aortic pressure and beat rate. The area of cavitation was found to decrease with increasing delay time at a constant aortic pressure, atrial pressure, and beat rate. Cavitation was found to occur in each of the fluids, with the most cavitation seen in the Newtonian fluids (distilled water and aqueous glycerin).

The Influence of Pressure Gradient on Desinent Cavitation From Isolated Surface Protrusions

An experimental investigation was conducted to study the desinent cavitation characteristics of various sizes of two-dimensional triangular and circular arc protrusions in a turbulent boundary layer for favorable, zero, and unfavorable pressure gradients. The roughness height (h ) varied from 0.025 cm (0.01 in.) to 0.762 cm (0.30 in.) and the relative height (h /δ) varied from 0.026 to 2.53. Desinent cavitation numbers (σd ) were obtained visually over a velocity range of 9.1 mps (30 fps) to 18.3 mps (60 fps) at an average total air content of 3.8 ppm (mole basis). The data for zero pressure gradient were in fair agreement with data obtained for the same protrusion shapes by Holl (1958). The cavitation number (σd ) was correlated with relative height (h /δ), Reynolds number (U δ/ν) and Clauser’s (1954) equilibrium boundary layer shape factor (G ) which includes the effect of pressure gradient. The data show that σd increases with pressure gradient. This result was not expected since it appears to contradict the trends implied by the so-called characteristic velocity theory developed by Holl (1958).

Real-time in vitro observation of cavitation on prosthetic heart valves

A method for real-time in vitro observation of cavitation on a prosthetic heart valve has been developed. Cavitation of four blood analog fluids (distilled water, aqueous glycerin, aqueous polyacrylamide, and aqueous xanthan gum) has been documented for a Medtronic/Hall prosthetic heart valve operating in a Penn State Electric Assist Device at physiologic conditions. For each fluid, cavity growth and collapse occurred in less than one millisecond. The cavity duration time was found to decrease with increasing atrial pressure at constant aortic pressure and beat rate. From this study, it was concluded that cavitation may also occur in blood.<<ETX>>

Hydrofoil Test Facility at ARL Penn State

The two-dimensional test section of the ARL/Penn State 12-inch water tunnel has been modified to allow a wide range of dynamic tests using hydrofoils. Three examples of test configurations for the hydrofoil test facility are given. These include tests of a single conventional hydrofoil with non-sinusoidal deflections, tests with two hydrofoils for studying tip vortex interactions, and tests of base ventilated supercavitating hydrofoils. When testing a single hydrofoil, the angle of attack can be varied as a function of time using a cam drive system. Arc length Reynolds number of over 1 million based on a 1.5-inch chord are possible. Hydrofoil lift, drag and pitching moment can be measured during transient operation with and without cavitation. Tip vortex interaction studies have been performed by using a second hydrofoil mounted upstream of the primary test hydrofoil. This upstream hydrofoil is inclined to the tunnel wall so only the tip projects in front of the downstream hydrofoil. The upstream hydrofoil can be traversed across the test section to study the tip vortex interactions. Supercavitating hydrofoils have been tested by ventilating behind a wedge installed along the tunnel wall upstream of the hydrofoil. A full range of test instrumentation is used to support the studies, such as LDV, PIV, high speed video, and acoustic measurements.Copyright

Specification of Acceptable Blade Porosity for Cavitation Performance of Marine Propellers

The issues of cavitation inception for porosity for cast propeller blades are addressed. Desinent cavitation measurements were made in the ARL Penn State 48-inch diameter water tunnel for a series of holes machined into inserts mounted in the tunnel test section. The data were compared to previous measurements with slots and isolated roughnesses. A design exercise is presented for a generic propeller blade to specify “acceptable” blade porosity for a given cavitation inception goal. The application to the blade accounts for the spanwise velocity distribution, local blade surface pressure distribution, local boundary layer thickness, and porosity size. The final result is a mapping of ranges in acceptable porosity for locations on the blade.Copyright