• Home »
• Science & Technology »
• Physics

This page details INL's Matched-Index-of-Refraction facility specifications such as flow facility design, flow quality, fluid selection, and facility temperature control and instrumentation.

Flow Facility Design

The MIR facility design incorporates the following design parameters:

• material compatibility with the refractive-index-matching fluid
• uniform temperature control to +/- 0.1 C
• flow rates up to about 340 liters/sec
• flow controllability down to less than ten percent of full flow
• modular system design
• system can be lengthened or shortened
• operable as a tunnel or open channel
• moderately low turbulence levels
• uniform cross-sectional velocity profile at the test section entrance
• negligible vibration.

Figure 1. The INL-designed MIR flow facility, fabricated by Idaho Steel Products Company in Idaho Falls, Idaho.

The material of construction for all components other than the test section is stainless steel. The direction of flow is clockwise in figure 1. The system components, from the test section moving clockwise, include expansion bellows, square cross-section elbow, square-to-round transition, expansion bellows, elbow, axial flow pump, bellows, round-to-square transition, expansion, elbow/flow distributor, settling chamber and contraction.

All expansion bellows are of stainless steel construction and include an inner sleeve, providing a smooth interior flow surface. These bellows are designed to minimize the transmission of vibrations to the test section as well as to accommodate thermal expansion. All elbows incorporate internal turning vanes to avoid flow separation or the formation of large secondary flow structures. The expansion downstream from the pump also includes vanes. These turning vanes are removable, if required.

The axial flow pump, manufactured by Aerolab Inc., is designed to provide a maximum flow rate of 340 liters/sec at about three meters (10 feet) water head and to operate with water or mineral oil as the working fluid. This maximum flow rate corresponds to a maximum test section velocity of 0.9 m/s. However, with an empty test section, velocities of about 1.7 m/sec have been measured at 500 rpm. The pump incorporates a viewing window to detect cavitation. The pump has four blades and is driven by a 75 hp motor in conjunction with a variable frequency controller.

Figure 2. Overview of the facility.

The flow distributor is designed to take flow from the rectangular bend and provide a uniform velocity distribution into the settling chamber. The distributor’s inlet cross-section is 0.61 x 1.22 m, while the outlet is 1.22 x 1.22 m, thus decreasing the average flow velocity by a factor of 2 between inlet and outlet. The intended design of the distributor is such that the confining surface approaches the outlet plane at an angle of about 63 degrees = tan-1(2/1) from the horizontal (or a convergence angle of 26.6 degrees). This 2:1 velocity ratio is maintained throughout the distributor by tapering the upper wall of the distributor at a constant 63 degree angle from the horizontal. This approach is a modification of the oblique header technique suggested by Professor A. L. London at Stanford University, and is used in water tunnels (London, Klopfer and Wolf, 1968; Humphreys and Reynolds, 1988).

The settling chamber has an interior cross section of 1.22 x 1.22 m and an overall length of 1.83 m. Within the settling chamber there are five slots for flow conditioning devices (e.g., honeycombs, screens, etc.). The settling chamber is made sufficiently oversized to allow the screen frames to be recessed into the wall. To provide a smooth interior wall, blank frames are fitted into slots not being used. Perforated surfaces at the top of the settling chamber allow for removal of trapped air through a bleed port at the top during filling.

Currently, a honey comb and three screens are employed. The honeycomb has 9.5 mm hexagonal cells, 76 mm deep. A 24 x 24 mesh, 0.19 mm wire diameter screen (porosity 67.2%) is installed in slot 2 just downstream of the honeycomb. Two finer screens, 30 x 30 mesh with 0.16 mm wire diameter (porosity 64.8%), are placed in slots 3 and 5. Also available is a 20 x 20 mesh, 0.22 mm wire diameter screen (porosity 67.2%). The honeycomb and all screens are fabricated from stainless steel. These devices can be used individually or in combination depending upon the flow characteristics required.

Figure 3. Test section and traversing mechanism with an external flow model (horizontal flat plate with roughness element) installed. The test section is operated as a closed channel.

The contraction has an area ratio of 4:1, reducing the 1.22 x 1.22 m cross section to 0.61 x 0.61 m. The contraction has zero slope and curvature at the inlet and outlet and 50% contraction at 50% of its length. The contraction is symmetrical in the horizontal and vertical planes. A short entrance section of stainless steel connects the polycarbonate test section to the contraction. This section is provided for installation of flow control devices such as grids, oscillating blades and/or transverse jets to increase turbulence levels and/or induce non-steady flows. The instrumentation ports in this section also permit insertion of pitot tubes and hot-film probes to measure entry profiles.

The test section (figures 3 and 4) is constructed of 3.8-cm thick polycarbonate plastic with an optical index-of-refraction of approximately 1.58. This thickness was chosen for structural strength and to minimize deflections under load. A stainless steel framework further stiffens the polycarbonate. The inside dimensions of the test section are 0.61 m wide x 0.61 m high x 2.44 m long. For access to the inside of the test section, the top wall is constructed of three removable sections. Modularity of the flow system allows replacement of this test section with another, such as a circular tube used for initial flow testing before this test section was fabricated. Although the polycarbonate is adequate for flow visualization, it is not of sufficient optical quality for precise LDA measurements. Therefore, window inserts of soda-lime-float-glass are provided at three locations on the two vertical walls of the test section. These windows are about 64 cm long and 28 cm high with a thickness of 1.9 cm.

The entire loop is supported on pneumatic vibration isolators. The vibration isolation system is installed to reduce the transmission of vibrations to the test section and to nearby instrumentation platforms. Also, the pneumatic system may be used to level the flow facility. Pressure regulators control the air pressure in the pneumatic bladders.

Figure 4. Test section with an internal flow model (ribbed annulus and converging annulus) installed. The test section is operated as an open channel.

Flow Quality

Vertical profiles of velocity were measured at the test section entrance for three fluid velocities and 1.2 m downstream from the entrance for a fluid velocity of 1.4 m/s. For these measurements, water was the working fluid. A pitot-static probe was used. At the entrance, the non-uniformity in velocity was about one percent of the mean velocity. At 1.2 m downstream, the non-uniformity in the velocity profile was less than 0.1 percent of the mean. Hot film measurements of velocity and turbulence were also measured. Again at a mean velocity of 1.4 m/s, the turbulence intensity was found to be less than 1 percent and dropped to less than 0.5 percent at velocities less than 1 m/s.

INL Facility Fluid Selection

Refractive-Index matching flow systems in the past have employed many different fluids or fluid mixtures, depending upon the refractive index being matched and the temperature range (Durst, Keck, and Kleine, 1979; Edwards and Dybbs, 1984; Budwig, 1994). For example a mixture of mineral seal oils was used to match glass, and mixtures of diesel oils as well as light fuel oil with dibutyl phthalate to match Pyrex glass. However, when designing large MIR facilities, more attention must be paid to issues such as fumes, toxicity, flammability and especially expense. These issues ruled out all the fluids used by past investigators and prompted a considerable investigation of the literature and discussion with vendors. Ultimately, INL engineers chose a light mineral oil, Penreco Drakeol #5. This fluid has the same index-of-refraction as fused quartz near room temperature, has no odor, is non-toxic, is relatively inflammable and non-volatile, and is inexpensive and very stable.

Viscosity, density and the refractive index of Drakeol #5 were measured as functions of temperature at the University of Idaho by Orr, Thomson and Budwig (1997) and Budwig (1998). The resulting property correlations are:

ρ (g/ml) = 0.8449 – 0.0005883 T

ν (cS) = 269.39 T ^–0.9366

η_D = 1.4664 – 0.0003587 T

with temperature in degrees Celsius. The thermal conductivity is about 0.098W/(mk) (Sparrell, 1995).

Facility Temperature Control and Instrumentation

Based on guidance from earlier studies, we estimated that the oil temperature must be controlled to be steady and uniform to within 0.3 C to match the refractive-index of quartz adequately for high-quality LDA measurements. Initially, installation of a heat exchanger(s) within the loop was considered. However, at low velocities the flow in the return loop would be expected to be laminar and the high Prandtl number (approximately 250) of the mineral oil implies that any thermal variation across the flow would decay very slowly with distance in the streamwise direction. Accordingly, a mixing approach utilizing jet injection is employed.

Approximately 300 l/min of the oil is extracted from the main loop through two pipes upstream of the main pump. Optimum temperature is achieved by means of a heat exchanger for cooling and a 10 DC kW heater (figure 1). In this approach, the heat exchanger provides coarse control and removal of thermal energy from the fluid and then the more precise variable power supply provides fine control to the heater to achieve the desired temperature. The fluid is reinjected into the main loop downstream from the main pump through 36 peripherally-located orifices of three mm diameters. This external temperature control system also includes a filter to remove any particulates of diameters greater than 5 µm. The remaining particulates are used as seeding.

Tests have been conducted to assess the performance of the temperature control system. The control system can maintain the fluid temperature to within ± 0.003 C of the specified temperature. Also at a pump speed of 400 rpm, fluid temperature of 29 C and ambient temperature of 24.6 C, the maximum temperature variation across the test section was measured to be 0.004 C.

A comparable temperature control system is used for the oil which flows through any internal flow models (e.g. Figure 4) and a turbulence generator employing injection.

Figure 5. Transmitting optics of two=component laser Doppler velocimeter mounted on three-directional traversing system (yellow).

Facility instrumentation can be divided into control/monitoring systems and for experimental measurements. Control instrumentation includes thermistor probes installed throughout the main loop, thermistors and a turbine flow meter in the model temperature control loop, data acquisition/control unit and laboratory computers. Various differential pressure transducers, pitot tubes and manometers have been used to monitor main loop flow rate at different locations. A hot-film anemometer system may be used to measure mean and rms velocities separately from the LDA system. National Instruments "LabView" software is used for data acquisition and control.

Velocity and turbulence measurements are primarily obtained with a two-component, TSI fiberoptic-based LDA operated in forward or back scatter mode (figure 5). When higher signal-to-noise ratios are necessary, custom-built receiving optics from the LSTM are used in forward scatter (figure 6). Although the backscatter LDV mode is more convenient in operation, the forward scatter signal-to-noise ratio is ten to 100 times higher, resulting in superior signal quality and validation rates, especially near an obstacle or wall. The test section was specifically designed with optical access on both sides such that LDA measurements in forward scatter could be obtained. With this configuration, the LDV measurement control volume diameter is approximately 60 µm. A traversing system is integrated with the LDV to control movement in all three axes.

Figure 6. Dr. Stefan Becker, from Lehrstuhl fur Stromungsmechanik, Universitat Erlangen, Germany, adjusts LDV receiving optics which he designed.

In order to match refractive-indices precisely, the refractive-index of the oil is varied by adjusting the oil temperature via the temperature control system. Since the index-of-refraction is wavelength-dependent, the optimum temperature for the green component of laser light is not quite the optimum for the blue. Thus, the temperature at which the indices are matched best for both components is determined experimentally. In one approach, the LDV is positioned so one of the two blue beams and one of the two green beams pass through a quartz obstacle (e.g., the blue through a roughness element and the green through a flat plate). The LDV data acquisition is operated in the two velocity coincidence mode for a series of short time series while the oil temperature is gradually varied. The temperature at which the maximum validated sampling rate is obtained serves to indicate the optimal operating conditions for this situation. For the oil in the INL MIR flow system this maximum sampling rate in coincidence mode conveniently occurs at about 23.7 to 23.8 degrees C.

References

1. Budwig, R. S., 1994. Refractive index matching methods for liquid flow investigations. Exp. Fluids, 17, pp. 350-355.
2. Budwig, R. S., 1998. Personal communication. Univ. Idaho, 15 January.
3. Durst, F., T. Keck and R. Kleine, 1979. Turbulence quantities and Reynolds stress in pipe flow of polymer solutions measured by two-channel laser-Doppler anemometry. Proc., 6th Symp. on Turbulence, Rolla, Mo.
4. Edwards, R. V., and A. Dybbs, 1984. Refractive index matching for velocity measurements in complex goemetries. TSI Quarterly, 10, No. 4, pp. 3-11.
5. Humphreys, W. W., and W. C. Reynolds, 1988. An experimental study of the effect of streamwise vortices on unsteady turbulent boundary-layer separation. Report No. TF-42, Thermosciences Div., Mech. Eng. Dept., Leland Stanford Junior Univ.
6. London, A. L., G. Klopfer and S. Wolf, 1968. Oblique headers for heat exchangers, J. Eng. Power, 90, pp. 271-286.
7. Orr, B., E. Thomson and R. S. Budwig, 1997. Drakeol 5 thermophysical property meausrements. Technical report, Mech. Engr. Dept., U. Idaho.
8. Sparrell, J. K., 1995. Letter, Sparrell Engineering Research Corporation, P.O. Box 130, Damariscotta, Maine 04543 USA, 14 March.