1、I n t ern at i on a 1 J o u rn al of H eat in g, Ven ti 1 a ti n g, 4 i r -C o n dit io II i n g a n d Ref ri get a t i n g Res e a tc II HVAC Vent. rate.* Velocity, YO Oh O? YO Location of Location of sampling point+ filter D:Under left hand F1 33 8 48 4 drawer of desk, 2 in. F2 70 18 78 10 (50 mm)
2、 from left F3 93 87 92 25 back comer F4 169 353 160 430 S: Midway between F1 80 13 83 13 shelves, 2 in. (50 F2 75 7 73 10 mm) from left comer F3 83 17 88 17 F4 72 11 73 21 L: In leg space F1 78 20 78 12 under desk, 2 in. F2 73 35 73 66 (50 mm) from top F3 99 75 101 69 - right back comer F4 87 114 -
3、G: Midway in gap F1 77 29 82 71 between desk and F2 79 61 78 67 wail, 12 in. (300 F3 105 141 103 1 O0 mm) above floor F4 93 43 94 68 G: As above, but F1 71 22 72 22 with horizontal F2 76 7 70 6 face of gap closed F3 95 156 97 137 F4 94 32 96 22 t See Figure 2. Relative to value at the room center wi
4、th the filter at F1: On high, q = 24.1 ach. u = 71.2 ft/min u36.2 cm/s). On low: a = 15.3 ach. u = 44.2 ft/min (22.5 cm/s). Again, the local ventilation rates did not vary as much as might be expected from the velocity, and were generally within about 3vo of the rate at the center of the room. The o
5、ne exception was under the left hand drawer (location D) with the filter at F1. where the ventilation rate was reduced by more than 50% with the filter on low and nearly 70% with the filter on high. The velocity at this location was no lower than at one or two other locations which showed only moder
6、ate reduction in ventilation rate. In cases where the filter was positioned such that the jet of filtered air impinged directly on the sampling location (for example location G with the filter at F3), the local ventilation rate was as high as or slightly exceeded that at the room center. The much hi
7、gher ventila- tion rates and velocities obtained at position D with the filter at position F4 are not sur- prising as the filter was within about 2 ft (610 mml of the measurement point. Closing the top face of the gap between the desk and the wail (location G) further indicated the lack of a correla
8、tion between velocity and ventilation rate. As shown in Table 2, closing the gap had a large effect on the velocity, yet the ventilation rates did not change much and were again within 30% of the room-center value. Of even more interest is that these findings suggest that effective ventilation of se
9、mi-enclosed regions can be achieved through a single opening. The dependence on the size of the opening and its orientation to the filter will now be examined in some detail. Ventiiation of an Enclosure To obtain further insight into the ventilation of partially enclosed regions, ventilation rates w
10、ere measured inside a box measuring 27 in. (686 mm) high, 23 in. (584 mm) wide and 28 in. (7 11 mm) deep with a volume of 10.1 ft3 (286 LI. Additional tests were done with the depth of the box reduced to 16.75 in. (426 mml and its volume to 6.0 ft3 ASHRAE TITLExIJHVAC 2-3 96 m 0759650 0523923 LTT m
11、182 HVAC - 0 30 30 20 20 10 10 O O c .- c ._ 3 O 1 2 3 40 1 2 3 4 O 1 2 3 40 1 2 3 4 Time, minutes Time, minutes Figure 8. Velocity measured at center of opening (circles) and near back of box (squares) MATHEMATICAL MODELING In modeling the ventilation of an enclosure with a single opening, the exch
12、ange with the surrounding air is generally considered to occur either as a result of turbulent eddies crossing the opening or by a pulsation type flow in and out of the enclosure (Cockroft and Robertson 1976. Holmes 1979, Haghighat et al. 1991, Narasaki et ai. 1991, van der Maas et al. 1991). Pulsat
13、ion dominates when the flow impinges directly on the opening (incident angle near O-), and eddy penetration dominates when the flow is parailel to the plane of the opening (incident angle near 90) (Haghighat et al. 1991). In our tests the dominant airflow near the box was parallel to the opening eve
14、n at an incident angle of O. This is because the jet from the filter stayed close to the floor until very close to the box and then turned up the face of the box. ASHRAE TITLE*IJHVAC 2-3 96 0757650 0523724 707 VOLUME 2, NUMBER 3. JULY 1996 185 Table 3. Velocities and Calculated Ventilation Rates for
15、 Box Diameter, d. Velocity through Calculated eddy Calculated ventilation inch (cm) opening, diffusivity Eq. IS), rate Eq. (4), ach ft/min (cm/s) R2/min (cm2/s) Small box, incident angle = O 10.0 (25.4) 32.5 (16.5) 9.0 (139) 58.9 7.5 (19. i) 33.1 (16.8) 6.9 (107) 33.7 6.0 (15.2) 33.1 (16.8) 5.5 (85)
16、 21.6 4.25 (10.8) 32.0 (16.3) 3.8 (59) 10.5 2.3 (5.9) 28.6 (14.5) 1.8 (28) 2.8 1.4 (3.6) 25.5 (13.0) 1.0 (16) 0.9 10.0 (25.4) 33.1 (16.8) 9.2 (142) 35.8 7.5 (19.1) 35.8 (18.2) 7.5 (1 16) 22.0 6.0 (15.2) 4.25 (10.8) 32.0 (16.3) 3.8 (59) 6.3 2.3 (5.9) 30.2 (15.3) 1.9 (30) 1.7 1.4 (3.6) 24.1 (12.2) 1.0
17、 (16) 0.5 Large box, incident angle = O Large box, incident angle = 45 10.0 (25.4) 31.9 (16.2) 8.9 (137) 34.6 7.5 (19. i) 31.0 (15.7) 6.5 (100) 18.9 6.0 (15.2) 4.25 (10.8) 28.3 (14.4) 3.4 (52) 5.5 2.3 (5.9) 21.7 (i 1.0) 1.4 (22) 1.3 1.4 (3.6) 21.9 (11.1) 0.8 (13) 0.5 Large box, incident angle = 90 1
18、0.0 (25.4) 17.3 (8.8) 4.8 (74) 18.8 7.5 (19.1) 16.4 (8.3) 3.4 (53) 10.0 6.0 (15.2) 15.4 (7.8) 2.5 (39) 6.0 4.25 (10.8) 13.7 (7.0) 1.6 (25) 2.7 2.3 (5.9) 10.7 (5.4) 0.7 (i i) 0.6 1.4 (3.6) 14.0 (7. i) 0.6 (9) 0.3 Large box, incident angle = 135 10.0 (25.4) 15.4 (7.8) 4.3 (66) 16.7 7.5 (19.1) 15.8 (8.
19、0) 3.3 (51) 9.6 6.0 (15.2) 12.2 (6.2) 2.0 (31) 4.8 4.25 (10.8) 11.9 (6.1) 1.4 (22) 2.3 2.3 (5.9) 9.9 (5.0) 0.7 (10) 0.6 1.4 (3.6) 7.4 (3.8) 0.3 (5) 0.2 Large box, incident angle = 180 10.0 (25.4) 15.8 (8.0) 4.4 (68) 17.1 7.5 (19.1) 16.4 (8.3) 3.4 (53) 10.0 6.0 (15.2) 16.1 (8.2) 2.7 (42) 6.3 4.25 (10
20、.8) 13.8 (7.0) 1.6 (25) 2.7 1.4 (3.6) 5.9 (3.0) 0.3 (4) o. 1 2.3 (5.9) 10.0 (5.1) 0.7 (10) 0.6 ASHRAE TITLE*IJHVAC 2-3 9b m 0359b50 0523925 845 m 186 HVACBrR RESEARCH Eddy Diffusion The air exchange is represented by an effective diffusion coefficient De If ce is the par- ticle concentration in an e
21、nclosure of volume V and c, the particle concentration in the room, then conservation of mass requires that for diffusion across an opening of area A where d is the characteristic length of the enclosure over which the concentration change takes place. Defining the ventilation rate of the enclosure
22、qe as Ade q =- e Vd Equation (3) becomes (4) subject to the initial condition ce = c, = ci at t = O. point to point and, using Equation (1). can be expressed as As discussed previously, the particle concentration in the room changes little from cr = ciexp(-q,t) (6) It follows from Equations (5) and
23、(6) that or To test Equation (8). some measure of the eddy diffusion coefficient De is required. It can be assumed that De = kd (9) where k is a constant, Y is the mean air velocity through the opening, and, following Holmes (1979). the characteristic length d can be taken to be the diameter of the
24、open- ing. The mean velocity measured at the center of the opening (see Table 3) was used to calculate the diffusion coefficient using Equation (9) with k = 0.33, and the ventilation rate for the enclosure could then be obtained from Equation (4). The calculated ventila- tion rates are included in T
25、able 3, and together with the known ventilation rate for the room (around 25 ach) were used to calculate concentration-time data for the box using Equation (8). The calculated concentrations for some of the geometries tested are shown as solid lines in Figure 6 and are in good agreement with the mea
26、sured concentrations. Effective ventilation rates for the box that can be compared directly with the experi- mentally determined values can be obtained using the calculated concentration data ASHRAE TITLEaIJHVAC 2-3 96 m 0759650 052392b 781 m VOLUME 2. NUMBER 3, JULY 1996 187 along with Equation (2)
27、. As shown in Figure 9, there was good agreement among the calculated and experimental data, confirming that the velocity through the opening is successfully correlated with the ventilation rates for ail the cases examined. We next examine how this velociw might arise from the flow originating at th
28、e filter unit. - _- - -_ 20 I Large box rn Small box rn 20 Experimentally determined ventilation rate, ach Figure 9. Comparison of experimental and calculated effective ventilation rates for box Pulsation Flow The air discharged from the filter creates a pressure difference across the opening as it
29、flows past or impinges against the enclosure. Fluctuations in the flow from the filter will result in fluctuations in the pressure difference driving the airflow through the opening. When the pressure just outside the opening exceeds that in the enclosure, air will flow into the enclosure and increa
30、se the pressure there. Due to the inertia of the moving air mass, the inside pressure may “overshoot,” causing air to flow back to the outside. If this process is repeated at high frequency, and provided the air inside the enclosure is well mixed, the pulsating flow will result in effective ventilat
31、ion. Following previous studies of wind ventilation through a single opening (Cockroft and Robertson 1976, Holmes 1979. Haghighat et al. 1991, Narasaki et al. 1991). thevelocity u of the airflow through the opening can be written in terms of an orifice equation as u = Co is the mean (root mean squar
32、e) of the pressure (relative to “ground“) just outside the opening. This pressure oscillates at the natural frequency of the enclosure ASHRAE TITLEmIJHVAC 2-3 96 m 0759650 0523928 554 m VOLUME 2. NUMBER 3, JULY 1996 189 Table 4. Calculated Parameters for Box Diameter of Area of Natural frequency, Hz
33、 Large box Small box opening, inch opening, Inertance cml cm2 kg/m4 -. 10.0 (25.4) 507 5.9 46.5 59.7 7.5 (19. i) 285 7.9 40.3 51.7 6.0 (15.2) 182 9.9 36.0 46.2 4.25 (10.8) 92 14.0 30.3 38.9 2.3 (5.9) 27 25.7 22.3 28.7 1.4 (3.6) 10 42.1 17.4 22.4 Capacitance-Large box: C = 2.0 x rn4.s2/kg4; Small box
34、: C = 1.2 x m4.s2/kg4. system (see Table 4).The discharge coefficient was assumed constant at C, = 0.6, although in the Reynolds number range prevailing (50 to 2500) the discharge coefficient can vary significantly, depending on the extent the flow is contracted as it enters the orifice. DISCUSSION
35、In general, there was little variation in the local ventilation rates in the room, includ- ing those at regions shielded from the filtered airflow. This uniformity in ventilation indicates that the air is well mixed throughout the room in spite of the substantial vari- ation in local velocities. It
36、seems that the high velocity air streams flowing from the filter rapidly entrain the surrounding air and so approach the average particle concentration within a short distance of the filter. The large variations in velocity found in shielded regions suggest that these regions might be cleaned by “pa
37、rcels” of air being periodically drawn into more rapidly flowing air streams. Although the mean velocity in the shielded regions is low, the rate at which the air here is exchanged with the more rapidly moving air streams must determine the ventilation rate. Two additional measurements seem to suppo
38、rt this premise: The velocity at a point just above the front of the desk top increased from about 38 ft/min to 75 ft/min (0.19 to 0.38 m/s) on moving the filter from position F1 to F2. With the filter in the first position, the measurement point was shielded from the filter by the desk and was prob
39、ably characterized by a wake or recirculating flow. On moving the fil- ter to F2 the measurement point was in direct line of sight of the filter. In spite of the doubling in velocity, the ventilation rate remained substantially the same (22.5 and 21.8 ach). However, if ventilation were to occur by p
40、lug flow, the velocity needed to achieve a ventilation rate of 24 ach in the room would be only about 4 ft/min (20 mm/s). This is about the same as the lowest mean velocity measured, and much lower than the 70 to 100 ft/min (0.35 to 0.5 m/s) measured at the center of the room. The data obtained with
41、 the box confirm that confined areas can be ventilated at acceptable rates as a result of a pulsation type flow through openings. From Equations (4) and (9) it is seen that the ventilation rate of a confined area is given by qe = ke V and so increases with the area A of the opening and decreases wit
42、h its volume V. Pro- vided the pulsations are at the natural frequency, the velocity through the opening is independent of these parameters (Equation 16). We recall that the measured velocity was indeed found to be fairly constant for the larger sized openings, although it did decrease for openings
43、less than about 5 in. (125 mm) in diameter. This decrease might ASHRAE TITLExIJHVAC 2-3 96 m 0759b50 0523929 490 m 190 HVAC there exist regions of flow separation and recirculation. Moreover, the heat exchanger contains numerous fins and tubes and, in addition, the fan contains a rotating element. T
44、his is further complicated by the humidity in the air, which means that the fluid may need to be specified in a two-phase form. As a result, investigations of flow phe- nomena in such systems have been largely confined to experimental tests carried out on the actual unit, or in a wind tunnel. Howeve
45、r, this is both costly and time-consuming. Moreover, it can only provide data for a limited number of configurations. Also, the numerous variables, which may influence the flow field, make the experiment difficult to perform. However computational fluid dynamics and heat transfer have now matured to
46、 the stage where state of the art commercial codes can address real engineering problems. However, the advances in finite volume numerical capability which permit this, such as full three-dimensionality, generalised Co-ordinate systems and multi-blocking, require convincing validation procedures to
47、demonstrate quantitative engineering potential. In this paper the computational fluid dynamics approach was used to determine the main characteristics of turbulent flow in a packaged air-conditioning unit. The method, which solves numerically a set of partial differential equations governing the flo
48、w, has been used extensively as a design tool in the aerospace industry. With proper validation, it can be used to examine many features of the problem quickly and effectively. As a result, the main objective of the present investigation was to establish a framework on which the modelling of turbule
49、nt recirculating flow within an air-conditioning unit can be based. The Finite Volume Method (FVM) was used to solve the Reynolds averaged Navier-Stokes equations. The turbulence in the flow was represented by the standard two-equation k- E model. A special modelling technique was developed for heat exchangers and centrifugal fans. To validate the model, an extensive experimental investigation was carried out in parailel, and detailed comparisons of theoretical and experimental results were made. SYSTEW DESCRIPTION The air-conditioning unit investigated was an in-l
