1、NA-04-6-2 Characterization of Airflows at the Exit of Registers Using Laser Doppler Velocimetry (LDV) Narinder K. Tutu, Ph.D. C.R. Krishna, Ph.D. Thomas A. Butcher, Ph.D. Member ASHRAE ABSTRACT This paper reports the results offluidflow measurements carried out at the exit of sidewall registers. Tim
2、e-averaged mean velocity and turbulence (local root mean square velocity fluctuation) profiles were measured at the exit plane of two commonly used 8 in. x 4 in. (203 mm x 102 mm) sidewall regis- ters. These data can then be used as an input boundary condi- tion in a computationalfluid dynamics (CFD
3、) code to predict the velocity and temperature distribution in an enclosure supplied by the registers. Laser Doppler velocimetry (LDV) was used to measure the axial and vertical components of the velocity vector at vari- ous locations across the face ofthe registers. Measurements af distances remove
4、dfrom the exitplane of the registers, but still within the “near field,” show how the complex profile due to the vanes transitions to ajet, and this canprovide apartial vali- dation measurement for the CFD results. Measurements were made at two diferent flow rates, and the evaluation of the results
5、suggests that the velocityfield at the exit ofboth of these registers scales with the flow rate through them. This means that, in the mode of operation in which the supplyfan (of an HVAC system) has a “high” and “low ”setting, similar veloc- ity scaling would result for these types of registers. INT
6、RODUCTION Air distribution systems are used to supply conditioned air to a physical enclosure whose environment needs to be regulated within certain limits. A key element of the air distri- bution system is the air supply outlet, which can be a register, a grille, or a diffuser. Its design determine
7、s the manner in which the supply air is distributed within the enclosure (to be simply referred to as a room hereafter) and the resultant mixing of the ambient room air with the supply air. In order to determine the ability of a given register design to produce an acceptable performance, detailed ve
8、locity and temperature measurements downstream of the register exit in the entire room must be performed. However, such a task is time inten- sive and costly. Consequently, the use of computational fluid dynamics (CFD) models to predict the airflow distribution in rooms has become necessary (see, fo
9、r example, Posner et al. 2003, Sorensen and Nielsen 2003, and Srebric and Chen 2002). In order to conserve the computational resources, and also because the pertinent choice of turbulence models for simulation of airflow within the register and within the room may be different (Sorenson and Nielsen
10、2003), the flow within the register is not normally simulated using the CFD (Srebric and Chen 2002). Instead, the flow through the register is simply prescribed as a boundary condition for the CFD model of the room. The prescribed flow through the register is either a simple model (such as the box m
11、ethod) or actual data from measurements. Our goal here is to follow the latter approach- that is, to perfom detailed velocity field measurements at the very exit ofthe register and then use these measurements as the boundary condition for the CFD code to predict the velocity and temperature distribu
12、tion within the room. The CFD predictions can then be used to determine the register perfor- mance parameters, for example, the “throw” as a function of flow rate. Some limited velocity and temperature measure- ments at selected downstream locations in the room can be performed to test the validity
13、of CFD predictions. Neverthe- less, the detailed near-field velocity measurements at the very Narinder K. Tutu, C.R. Krishna, and Thomas A. Butcher are mechanical engineers on the scientific staff of Brookhaven National Labo- ratory (BNL), Upton, N.Y. Thomas A. Butcher is also the head of the Energy
14、 Resources Division in the Energy Sciences and Technology Department at BNL. 02004 ASHRAE. 709 exit of the register are sufficient to Characterize the fluid dynamic performance of the register design. During this investigation, the near-field velocity measure- ments at the exit oftwo different comme
15、rcially available regis- ters were performed. The identity of the registers, designated here simply by the letters A and B, is given in the report by Tutu et al. (2003). Register A (shown on the bottom portion of Figure 3) has an opening of 187 mm x 87 mm. The hydraulic diameter, D, of the register
16、opening based upon these dimen- sions is calculated to be 0.1 19 m. It was oriented with the longer dimension in the vertical direction. It has a set of 13 flat fixed vanes. The central vane is about 12.5 mm wide and oriented normal to the flow. The top set of six vanes (and the bounding top edge of
17、 the register opening) point upward and are oriented at an angle of 30+2 degrees to the horizontal, while the bottom set of six vanes (and the bounding bottom edge of the register opening) point downward and are oriented at an angle of 30+2 degrees to the horizontal. The spacing between the vanes is
18、 about 12.2 mm. Register B (shown in Figure 8) has an opening of 189 mm x 87 mm. The hydraulic diameter, D, of the register opening based upon these dimen- sions is calculated to be O. 119 m. It was oriented with the longer dimension in the horizontal direction. It has a set of three curved vanes th
19、at can be rotated independently along a horizontal axis parallel to the vane length. The spacing between the vanes is about 25 mm, and the vanes are about 27 mm wide. A laser Doppler velocimeter (LDV) was used to measure two components of the velocity vector at the exit of two regis- ters at various
20、 locations across the face of the registers. Details . As discussed later, the velocity field at the exit of the register is influenced by the design of the “boot”-the transition piece between the register and the supply duct. Thus, for practical appli- cations the boot and the register should be co
21、nsidered as an inte- grated system. of the experimental facility and instrumentation are presented in the next section. This is followed by the measurements for the registers A and B. It is emphasized here that the intent of these measurements is not to evaluate the register performance directly in
22、a typical home setting (where the register is likely to be located close to the floor or ceiling) but to provide detailed near-field measurements at the register exit for use in a CFD simulation. EXPERIMENTAL SETUP AND MEASUREMENT TECHNIQUE Description In order to ensure a well-characterized flow wi
23、thout any anomalies into the register, a flow conditioning system was designed and constructed. Although the duct systems used in most American homes and small businesses may not have uniform and symmetric incoming flow, it is nevertheless important to ensure that the airflow entering into the regis
24、ter- boot combination is relatively uniform and symmetric. Other- wise, any unanticipated or interesting observations will always be suspected as having been probably caused by the peculiarities of the flow delivery system. For this reason, this flow field (spatial velocity profile at the entrance t
25、o boot) must be shown to be symmetric and without anomalies by measure- ment prior to testing of the registers. A schematic of the flow conditioning system and the experimental apparatus is shown in Figure 1. A variable speed 254-mm-diameter duct blaster was used as a source for the air delivery. As
26、 shown in Figure 1, it delivers air to the flow conditioning system, which consists of a “diffuser,” followed sequentially by a 1.2-m-long and 457-mm-diameter settling chamber, a contraction, and a 1 .5-m-long, 178-mm-diameter duct. The “settling chamber” has a “flow straightener” (denoted by EC in
27、Figure i), which consists of 152-mm-deep and 25.4-mm-square openings with - . _ RETURN DUCT “= _ . . P - - - BURST PROCESSOR CONTROLLER - - IF- -WALL tj * i 1 TESTROOM 1 EUPfENTAND RGWCON)ITIONINGROOM - 3.81m 6.60m * I PLAN (TOP VIEW) PI PARTICLE INJECTION. (6.SSmm 10 Metal tube injects particles al
28、ong the centerline Of plpe) EC 162mm Deep ”Egg Crate”with 26.4mm x Z6.4mm openings and e screen on either side The Center Line of the flow system is at an elevation of 1.6m above the floor of the Test Room Heightof the Test Room: 2.69m. Figure I A schematic of the experimental facility for the study
29、 of velocity projles ut the exit of HVAC registers. Xand Y Ure the axial und lateral coordinates, respectively. 71 O ASHRAE Transactions: Symposia a screen on either end. The purpose of the flow straightener is to remove any swirling components of velocity from the airflow, and also to dampen any fl
30、ow nonuniformities. Since the laser Doppler technique uses the scattered light from particles in the flow to measure the flow velocities, parti- cles need to be injected into the flow. A commercial “fluidized bed particle generator” was used for this purpose to inject tita- nium dioxide particles in
31、 the airstream. The particles were injected along the centerline of the 178-mm-diameter duct via a 6.35 mm internal diameter tube. The particle injection loca- tion is shown in Figure l. The nominal mean diameter of these particles is 3 microns. The LDV system consists ofthe follow- ing components:
32、a. Laser, detector, and optics module (designated in the text by “LDV probe”) b. Digital burst processor The laser has a wavelength of 685 nm and a power of 50 mW (25 mW for each of the two beams). The burst proces- sor contains the electronics to process the signals from the LDV probe. It is contro
33、lled from a desktop personal computer (PC) via proprietary LDV software, and the processed data are transferred to the PC. A 350-mm focal-length lens is connected to the front end of the LDV probe. It outputs two laser beams that are 50 mm apart at the exit of the lens. To enable the measurement of
34、the direction of the velocity component, the frequency of one of the beams is shifted by 40MHz by the Bragg cell in its path. The measurement volume (the location at which the instantaneous velocity is measured) formed by the region of intersection of the two laser beams is an elongated ellipsoid th
35、at is 3.8 mm in length, with minor axes of 0.3 and 0.1 mm. For ail the measurements performed here, the 3.8-mm-long major axis was always aligned parallel to the lateral Y-coordinate direction. The velocity data from a given location consist of a series of instan- taneous values: U, U, U3. U, which
36、are recorded at discrete times tl, t, t3. t,. These data can then be used to calculate such items as the time-averaged mean velocity and the root mean square velocity fluctuation. During the experiments, the LDV software was set up to collect 1500 velocity samples for a maximum duration of 90 second
37、s at each measurement loca- tion. Toward the edges of the flow field, where the particle concentration and the flow velocities are low, the total number of data points collected during the 90-second interval was much less than 1500. For the instrument settings used during these experiments, the unce
38、rtainty in the measurement of instantaneous velocity of individual data points, Ui, is esti- mated to be 0.006, ms. The uncertainty (standard deviation) in the computed temporal mean velocity component due to the *. The laser Doppler velocimeter computes the instantaneous veloc- ity U from the relat
39、ion U = dfF, where dfis the fringe spacing in the measurement volume (provided via calibration by the manu- facturer) and F is the measured doppler frequency. Since the uncertainties in df and F are statistically independent and are known from the instrument specifications, the uncertainty in U is e
40、asily calculated. finite sample size (N= 1500) is given by (ulfi), where u is the root mean square fluctuation in the velocity component being measured. Thus, for a location where u = 1 m/s, the total uncertainty (standard deviation) in the measured temporal mean velocity component is estimated to b
41、e 0.03 ms. The LDV probe was mounted on a motor-controlled three-axis traversing mechanism. The traverse has a total travel of 245 mm in each of the two horizontal directions, and a travel of 375 mm in the vertical direction. The position ofthe LDV probe is indicated on the front panel display of th
42、e traverse controller with a resolution of 0.005 mm. This posi- tion is also transferred to the PC and recorded along with the velocity data. By installing a 90-degree bracket on the vertical traversing axis of the traversing mechanism, the LDV probe can be rotated 90 degrees, which effectively rota
43、tes the plane formed by the two laser beams leaving the LDV probe. Since the LDV measures the component of velocity along the plane of the laser beams and normal to a line that bisects the angle between the two beams, two orthogonal velocity components could be measured by rotating the LDV probe by
44、90 degrees. During these experiments, the axial (U) and the vertical ( W) velocity components were measured. Although all the three velocity components are needed for the complete character- ization of the velocity field, only two components of the veloc- ity vector were measured. However, the chose
45、n register designs were such that the velocity vectors at the exit of these registers have only two dominant velocity components. Excluding the outer edges of the flow, the temporal mean value of the third velocity component (which was not measured) is expected to be very much smaller than the axial
46、 component. Therefore, and also because of time constraints, the third velocity component was not measured. Velocity Profiles at the Exit of the Flow Conditioning System The registers to be tested were 203 mm x 102 mm in cross section. These are typically screwed onto a 203 mm x 102 mm boot. The boo
47、t, in turn, is connected to the air delivery system via a 102-mm-diameter duct. As shown in Figure 1, a short length of 102 mm (4 in.) diameter sheet metal duct was connected to the end of the flow conditioning system. The flow conditioning system was then moved until the end of this duct was flush
48、with the location where the register would be located during the tests. The LDV instrumentation was then set up to measure velocity profiles at the exit of this duct. Measurements of the axial velocity component across the duct exit along diameters at four different angles relative to the horizontal
49、 and at a distance of 20 mm from the duct exit were performed. Both the temporal mean and the turbulence level in the axial velociy were computed. These are plotted in Figure 2. As can be seen, the velocity profile is reasonably uniform and symmetric. Furthermore, the turbulence level is low. These measurements show no anomalies and demonstrate that the flow conditioning system is working as designed. The nominal flow rate during these measurements was 0.0387 m3/s ASHRAE Transactions: Symposia 71 1 124.8 to 126.5 cfm. The nominal value ofthe airflow rate at
