Experimental Apparatus And Procedures

The overall experimental apparatus (but not the test section) is similar to the one described by Mahmood et al. [37]. A brief description of this apparatus is also presented here.

Channel and test surface for heat transfer measurements. A schematic of the facility used for heat transfer measurements is shown in Fig. 2.1a. The air used within the facility is circulated in a closed-loop. One of three circuits is employed, depending upon the Reynolds number and flow rate requirements in the test section. For Reynolds numbers ReH less than 10,000, a 102 mm pipe is connected to the intake of an ILG Industries 10P type centrifugal blower. For ReH between 10,000 and 20,000, the same pipe is connected to the intake of a Dayton 7C447 1.0 hp centrifugal blower. For higher Reynolds numbers, a 203 mm pipe is employed with a New York Blower Co. 7.5 HP, size 1808 pressure blower. In each case, the air mass flow rate from the test section is measured (upstream of which ever blower is employed)

using an ASME standard orifice plate and Validyne M10 digital pressure manometer. The blower then exits into a series of two plenums (0.9 m square and 0.75 m square). A Bonneville cross-flow heat exchanger is located between two of these plenums, and is cooled with liquid nitrogen at flow rate appropriate to give the desired air temperature at the exit of the heat exchanger. As the air exits the heat exchanger, it enters the second plenum, from which the air passes into a rectangular bell mouth inlet, followed by a honeycomb, two screens, and a two-dimensional nozzle with a contraction ratio of 5.6. This nozzles leads to a rectangular cross-section, 411 mm by 103 mm inlet duct which is 1219 mm in length. This is equivalent to 7.4 hydraulic diameters (where hydraulic diameter is 164.7 mm). Two trips are employed on the top and bottom surfaces of the inlet duct, just upstream of the test section, which follows with the same cross-section dimensions. It exits to a 0.60 m square plenum, which is followed by two pipes, each containing an orifice plate, mentioned earlier.

Figure 2.2 gives the geometric details of the test surface, including rib turbulator geometry. A total of 13 ribs or rib segments are used on the top wall and on the bottom wall of the test section. As mentioned, these are arranged with 45 ° angles with respect to the streamwise flow direction, such that the ribs on opposite walls of the channel are perpendicular to each other. Each rib has 12.8 mm height and square cross-section. The ratio of rib height to hydraulic diameter is .078, the rib pitch-to-height ratio is 10, and the blockage provided by the ribs is 25 percent of the channel cross-sectional area. The top wall of the test section also has two cut-out regions (one at the upstream end and one at the downstream end) where a zinc-selenide window can be installed to allow the infrared camera to view a portion of the test surface on the bottom wall. When this window is not in use, inserts with ribs (which exactly match the adjacent rib turbulators on the top wall) are used in its place. Also identified in Fig. 2.2 is the test section coordinate system employed for the study. Note that the y coordinate is directed normal to the bottom wall.

All exterior surfaces of the facility (between the heat exchanger and test section) are insulated with Styrofoam (k =0.024 W/mK), or 2 to 3 layers of 2.54 cm thick, Elastomer Products black neoprene foam insulation (k =0.038 W/mK) to minimize heat losses. Calibrated copper-constantan thermocouples are located between the three layers of insulation located beneath the test section to determine conduction losses. Between the first layer and the 3.2 mm thick acrylic test surfaces are custom-made Electrofilm etched-foil heaters (each encapsulated between two thin layers of Kapton) to provide a constant heat flux boundary condition on the test surface. The acrylic surfaces, which are adjacent to the airstream, contain 35 copper-constantan thermocouples, which are placed within the ribs as well as within the flat portions of the test surface between the ribs. Each of these thermocouples is located 0.0508 cm just below this surface to provide measurements of local surface temperatures, after correction for thermal contact resistance and temperature drop through the 0.0508 cm thickness of acrylic. Acrylic is chosen because of its low thermal conductivity ( k =0.16 W/mK at 20 ° C) to minimize streamwise and spanwise conduction along the test surface, and thus, minimize "smearing" of spatially varying temperature gradients along the test surface. The power to the foil heater is controlled and regulated using a variac power supply. Energy balances, performed on the heated test surface, then allow determination of local magnitudes of the convective heat flux.

The mixed-mean stagnation temperature of the air entering the test section is measured using five calibrated copper-constantan thermocouples spread across its cross-section. To determine this temperature, thermocouple-measured temperatures are corrected for thermocouple wire conduction losses, channel velocity variations, as well as for the differences between stagnation and recovery temperature. All measurements are obtained when the test facility at steady-state, achieved when each of the temperatures from the thermocouples (on the bottom test surface) vary by less than 0.1 ° C over a 10 minute period.

Local Nusselt number measurement. To determine the surface heat flux (used to calculate heat transfer coefficients and local Nusselt numbers), the convective power levels provided by the etched foil heaters are divided by flat test surface areas. Spatially-resolved temperature distributions along the rib turbulator test surface are determined using infrared imaging in conjunction with thermocouples, energy balances, digital image processing, and in situ calibration procedures. To accomplish this, the infrared radiation emitted by the heated interior surface of the channel is captured using a VideoTherm 340 Infrared Imaging Camera, which operates at infrared wave lengths from 8 m m to 14 m m. Temperatures, measured using the calibrated, copper-constantan thermocouples distributed along the test surface adjacent to the flow, are used to perform the in situ calibrations simultaneously as the radiation contours from surface temperature variations are recorded.

This is accomplished as the camera views the test surface through a custom-made, zinc-selenide window (which transmits infrared wave lengths between 6 and 17 m m) located on the top wall of the test section. Reflection and radiation from surrounding laboratory sources are minimized using an opaque shield which covers the camera lens and the zinc selenide window. Frost build-up on the outside of the window is eliminated using a small heated air stream. The window is located either just above portions of the second, third, and fourth ribs, or just above portions of the tenth, eleventh, and twelfth ribs downstream from the leading edge of the test surface. Eleven to thirteen thermocouple junction locations are usually present in the infrared field viewed by the camera. The exact spatial locations and pixel locations of these thermocouple junctions and the coordinates of a 12.7 cm by 12.7 cm field of view are known from calibration maps obtained prior to measurements. During this procedure, the camera is focused, and rigidly mounted and oriented relative to the test surface in the same way as when radiation contours are recorded.

With these data, gray scale values at pixel locations within video taped images from the infrared imaging camera are readily converted to temperatures. Because such calibration data depend strongly on camera adjustment, the same brightness, contrast, and aperture camera settings are used to obtain the experimental data. The in situ calibration approach rigorously and accurately accounts for these variations.

Images from the infrared camera are recorded as 8-bit gray scale images on commercial videotape using a Panasonic AG-1960 video recorder. Images are then digitized using NIH Image v1.60 software, operated on a Power Macintosh 7500 PC computer. Subsequent software is used to convert each of 256 possible gray scale values to local Nusselt number values at each pixel location using calibration data. Thermal conductivity in the Nusselt number is based on the average of the local wall temperature and the temperature of the air at the upstream inlet. Contour plots of local surface temperature and Nusselt number are prepared using DeltaGraph v4.0 software. Each individual image covers a 300 pixel by 300 pixel area. Sargent et al. [38], and Mahmood et al. [37] provide additional details on the infrared imaging and measurement procedures.

Friction factors. Wall static pressures are measured along the test section simultaneously as the heat transfer measurements are conducted, using 12 side wall pressure taps, located 25.4

mm apart near the downstream portion of the test section. These measurements are made in the test section with rib turbulators, as well as in a baseline test section with smooth surfaces on all four walls. Friction factors are then determined from streamwise pressure gradient magnitudes. Pressures from the wall pressure taps are measured using Celesco LCVR pressure transducers. Signals from these transducers are processed using Celesco CD10D Carrier-Demodulators. Voltages from the Carrier-Demodulators are acquired using a Hewlett-Packard 44422A data acquisition card installed in a Hewlett-Packard 3497A data acquisition control unit, which is controlled by a Hewlett-Packard A4190A Series computer

Time-averaged flow velocity components and pressure. A separate channel facility (but with the same test section), and the same interior geometry identical to that in the heat transfer facility, is employed for flow visualization as well as quantitative surveys of flow structure. This facility is shown schematically in Fig. 2.1b.

A 1.27 mm diameter miniature five-hole pressure probe, manufactured at the University of Utah especially for these measurements, is used to obtain time-averaged surveys of total pressure, static pressure, and the three mean velocity components. These data are then used to deduce distributions of streamwise vorticity. To obtain the surveys, the probe employed is mounted on an automated two-dimensional traverse, and inserted into the test section through a slot lined with foam to prevent air leakage. The output ports of the probe are connected either to Validyne DP 103-06 pressure transducers (to measure differential pressures up to 2.5 mm of water), or Celesco LCVR pressure transducers (to measure differential pressures up to 20.0 mm of water). Signals from the transducers are then processed using Celesco CD10D Carrier-Demodulators. Voltages from the Carrier-Demodulators are acquired using a Hewlett-Packard

44422A data acquisition card installed in a Hewlett-Packard 3497A data acquisition control unit. This control unit, the Superior Electric type M092-FD310 Mitas stepping motors on the two-dimensional traverse, a Superior Electric Modulynx Mitas type PMS085-C2AR controller, and a Superior Electric Modulynx Mitas type PMS085-D050 motor drive are controlled by a Hewlett-Packard A4190A Series computer. Contour plots of measured quantities are generated using a polynomial interpolating technique (within DeltaGraph software) between data points. In each survey plane, 1560 data points are spaced 2.54 mm apart. Additional details of the five-hole probe measurement procedures, including calibration details, are given by Ligrani et al. [39,40].

Flow visualization. Flow visualization using smoke is used to identify vortex structures and other secondary flow features. Smoke from four horizontally-oriented smoke wires is employed for this purpose. These are located 4.8 mm, 11.9 mm, 88.5 mm, and 95.7 mm from the bottom test surface 25 to 29 mm from the downstream edge of the test section, which is equivalent to X =1258-1261 mm. To accomplish this, each wire is first coated with Barts Pneumatics Corp. super smoke fluid and then powered using a Hewlett-Packard 6433B DC power supply. With this arrangement, the smoke forms into single thin lines parallel to the test surface. As the smoke is advected downstream, the secondary flows which accompany vortex and secondary flow development cause the smoke to be rearranged in patterns which show the locations and distributions of these flow phenomena. Smoke patterns are illuminated in a spanwise-normal light plane located at X =1462 mm using a thin sheet of light provided by a Colotran ellipsoidal No. 550, 1000 watt spot light, and a slits machined in two parallel metal plates. Images are recorded using a Panasonic WV-BP330 CCTV video camera, connected to a Panasonic AG-1960 type 4-head, multiplex video cassette recorder. Images recorded on video tape (taken individually or in sequence) are then digitized using a Sony DCR-TRV900 digital video camera recorder. The resulting images are then further processed using a Power Macintosh 7500 PC computer, and finally printed using a Panasonic PV-PD 2000 digital photo printer. Additional discussion of the procedures used for flow visualization is provided by Ligrani [41].

0 0

Post a comment