1、AN-04-4-3 Measurement of Pore Size Variation and Its Effect on Energy Wheel Performance Wei Shang, Ph.D. Robert W. Besant, P.Eng. Fellow ASHRAE ABSTRACT Two experimental techniques are used to investigate variations in airflow channel geometry for four energy wheels that transfer heat and water vapo
2、r between exhaust and supply airflows. Using optical and micrometer measure- ments andpressure drop probe measurements, the flow chan- nel hydraulic diameters are shown to J;t Gaussian distributions and the ratio of standard deviation of hydraulic diameter to mean hydraulic diameter, o/D, , is calcu
3、lated for a parallel surface, honeycomb, and two dierent corru- gated airflow channel matrices for four energy wheels. The optical and micrometer results for these wheels showed a range of o/D, from 0.064 to 0.234, indicating a range of pressure drop divided by the pressure drop with no variation (i
4、e., o = O), Ap/Apo,from 0.982 to 0.859 and effectiveness ratio, E/, from 0.986 to 0.867 for the wheels tested. The pressure probe results show a small precision uncertain, but, due to important bias errors, pressure probe results must be calibrated using the optical and micrometer results. INTRODUCT
5、ION Regenerative wheel pore sizes can vary due to slight vari- ations in manufacturing control setpoints for the assembly of matrices for regenerative wheels. Other than the papers by London (1970) and Mondt (1 977), which implied some flow channel pore size variations, there is no reported physical
6、 evidence of these variations even though manufacturers gener- ally know that the flow channel size variations should be kept small. While Shah and London (1 980) showed that a reduction in channel size variations results in a increase in pressure drop ratio, ApAp, and the variation in effective hea
7、t transfer coef- ficients, they did not consider how the effectiveness of a wheel would change. In the recent paper by Shang and Besant (2003), the problem of random pore size variations among the flow channels for regenerative exchangers was investigated theoretically for random variations in pore
8、sizes. It was shown that for random variations in the flow channel pore sizes the pressure drop across these exchangers and effectiveness (sensible, moisture or latent, total or enthalpy) decrease rela- tive to the same wheel with the same total mass flow rate with a uniform pore size. The pore size
9、 variation in a wheel can be measured directly using an optical magnification and micrometer system or indirectly using a pressure drop probe. In this paper these methods are used to investigate pore size variations in typical desiccant-coated energy wheels that are used to exchange heat and moistur
10、e between ventilation supply and exhaust air flows. These measured flow channel variations are then used with the theoretical results to infer the expected decreases in pressure drop and effectiveness. Four different energy wheel matrices were investigated to illustrate the method of testing and ana
11、lysis-one with parallel channel surfaces, one with honeycomb pores, and two with corrugated pores. These wheels were selected for their differ- ent flow channel or pore shapes and materials of consruction and are not expected to be representative of a particular manu- facturing method. Typical optic
12、al images are shown in Figure 1 for each matrix. The overall geometric properties of these wheels are listed in Table 1. The characteristic dimensions for each of these types of flow channel pore are presented below each of the photos for the four types of matrices. The dotted lines in the Figure 1
13、b schematic of a honeycomb represent the length of lines connecting the opposite vertices. The sum of these is propor- tional to the flow channel hydraulic diameter when small vari- ations occur. Wei Shang and Robert W. Besant are with the Department of Mechanical Engineering, University of Saskatch
14、ewan, Saskatoon, Canada. 41 O 02004 ASHRAE. Table 1. Geometric Properties for Different Regenerative Wheels Energy Wheel 1 2 Pore Material Pore Shape Hub Diameter (mm) Wheel Thickness (mm) Plastic Parallel 130 f 0.5 38 f 0.1 Paper Honeycomb 55 f 0.5 101 f 0.5 3 4 (a) Parallel surface wheel I (b) Hex
15、agonal honeycomb Aluminum Corrugated 80 f. 0.5 99 f 0.5 Paper Corrugated 104 f 0.5 101 f 0.5 (c) Corrugated aluminum wheel m (d) Corrugated paper wheel 2a - It Figure 1 Photos of the matrices for four different energy wheels. PRESSURE DROP AND EFFECTIVENESS RATIOS In the paper by Shang and Besant (2
16、003), pressure drop ratio, Ap/Ap, and effectiveness ratio, (b) hexagonal honqcomb wheel (D); (e) corrugated aluminum wheel (2b); (d) corrugatedpaper wheel (2b). random or Gaussian for the 60 random positions selected for each wheel (Taylor 1982). That is, x2 I 6 for each wheel. Figure 8 shows the ch
17、i squared analysis of thc characteristic dimension distributions of the four wheels where the bars give the data and the lines give the Gaussian distribution. For the corrugated wheels, the variation in the wave length, 2a, was always much smaller than the variation in the wave amplitude, b, so the
18、prime reason for variation in the hydraulic diameter, D, is due to variations in the wave amplitude, b, as shown in Figure 8. Parallel Surface Matrix The parallel surface flow channel matrix used dimples circumferentially spaced 13 mm apart on every second plastic sheet to obtain the spacing between
19、 the sheets when they are placed in a wheel. Some of these dimples are clearly evident in Figure la for the parallel surface matrix. Figure 6 shows the micrometer data for the thickness and dimple measurement of the plastic sheet with the dimples where the measurements were taken using a micrometer
20、with an uncertainty off 0.003 mm (ASME 1985). These data show a mean value for the thickness of various sheets to be O. 162 mm and a mean value for the sheet thickness plus dimple height to be 0.432 mm. While this thickness has a standard deviation of less than 0.004 mm, the standard deviation of th
21、e thickness plus sheet thick- ness is 0.020 mm. These dimple height variations imply vari- ations in the spacing between the dimpled sheets. The adjacent smooth plastic sheet was measured to be somewhat thicker than the dimpled sheet; it has an average height of 0.229 mm with a standard deviation of
22、 0.007 mm. The optical spacing measurements for the sheets in the wheel are shown in Figure 7. These data do not include the bias caused by the fact that some of the airflow channels are not exactly parallel to the optical light beam. This bias was investigated by optically measuring the total dista
23、nce between 40 flow channels in the wheel at several locations on the wheel. It was found that the average bias was 0.040f0.001 nun for each flow channel, implying that the flow channel width on average is 0.040 mm greater than the optical measurements of channel width. That is, the bias-corrected f
24、low channel width is equal to the value in Figure 7 plus 0.040 mm. From these corrected data we can now calculate the standard deviation of the flow channel width divided by the average channel width, which is equal to the standard deviation of hydraulic diameters divided by the mean hydraulic diame
25、ter (OX),). For this wheel the value of this dimensionless term is 0.23410.030 or 23.4%+3.0%. The uncertainty in this calculated ratio is calculated using the precision uncertainty of the distance data. Knowing the ratio of standard deviation of pore hydraulic diameters to its mean value and assumin
26、g that the dimples will have a negligible impact on the pressure drop or effectiveness for laminar airflow through the channels, the ratios ofpressure drop and effectiveness, as presented by Shang and Besant (2003), can be calculated. The corresponding values are Ap/ Apo = 0.859 or 85.9% and E/ howe
27、ver, it aver- ages these flow channel pore properties for the probe diameter used. So the pressure probe can only give variations in the flow characteristics when the pore size variations are not exactly uniformly distributed on the wheel face. Therefore the pres- sure probe results must be calibrat
28、ed with the optical results for each type of wheel. The utility of the pressure probe method might be to do either low-cost on-line checking of variations in pore sizes during the manufacturing or identifying energy wheels in HVAC applications with suspect or apparent large pore size variations due
29、to manufacturing or on-site damage. On-site use of the opticallmicrometer method would likely prove more difficult and expensive. Test Facility Figure 9 shows the schematic of pressure drop probe test facility for pore channel size variation measurement on an energy wheel. This system contains sever
30、al components-a compressed air supply, a calibrated mass flow meter, a cali- brated flow meter controller, and a calibrated pressure trans- ducer. In operation this system supplies a constant mass flow of air through the probe. The flexible inner diameter of plastic tube connecting the probe to mass
31、 flow meter was 8.0f0.1 mm. This tube is inserted through a solid rubber disk, which can be placed on a flat surface with no air leakage except through any pores below the 8 mm tube. A data acquisition system comprised the flow meter output, pressure transducer output, and a computer that records th
32、e flow rate and pressure 41 7 Table 4. Bias and Precision Uncertainties for the Pressure Drop Measurement Bias Uncertainty Precision Uncertainty f 2 Pa (1 0.008 in. H20) Pressure transducer f 3 Pa (I 0.012 in. H20) Mass flow meter +O PRESSURE DROP MEASUREMENT ANALYSIS AND U N C E RTAI N TY .o01 gls
33、For laminar flow through the pores, the equation for the pressure drop across the probe is the same as the pressure drop across the wheel for air flow through wheel matrix in an HVAC application. This equation, which is a slight modification to the equation for fully developed laminar flow through c
34、han- nels of any cylindrical shape (Shah and London 1980), is (Shang and Besant 2003): * 0.003 gls where Kf = friction constant, p = viscosity of air (Pas), L = thickness of wheel (m), p = density of air (kg/m3), L = thickness of wheel (m), qp Y = porosity of the wheel under the probe, Ap d = averag
35、e hydraulic diameter of pores under the probe From this equation we can get the ratio of pore diameters between any two positions on the wheel surface, Pl(i, j) and however, if the probe diameter is very large then we cannot expect to get large variations in the probe pressures because the range of
36、pore sizes under the probe increases with probe size. Indeed, the probe at any position on a wheel aver- ages the pressure drop for the pores under the pressure probe. For a given wheel with a random variation in pore sizes over the entire wheel, the pores under the probe placed at a large number of
37、 positions can be expected to result in a pressure drop distribution corresponding to a narrower range of average pore sizes than would be the case if individual pores were opti- cally measured. Consequently, the standard deviation of the probe pressure drop data using Equation 11 will always be sma
38、ller than the standard deviation of optical diameter data. The amount of decrease in this standard deviation using pres- sure drop data compared to optical and micrometer data depends on the manufacturing method used for the pores in the wheel. 41 8 ASHRAE Transactions: Symposia Table 5. Comparison
39、of the Optical-Micrometer and Pressure Probe Results for Energy Wheels Wheel Matrix Parallel surface Honeycomb _ Optical Pressure Probe 23.4% f 3.0% NIA 7.6% f 1.0% 1.75% f 0.44% Aluminum corrugated _ 6.4% k 2.8% 2.52% f 0.30% L Paper corrugated Pressure Drop Measurements Figures 5 and 9 show the pr
40、obe at position Pu on the face of an energy wheel. This probe is manually positioned on a wheel face for several radial positions and angles. In this study 10 radial probe positions (50,70,90, 110, 130, 150, 170, 190, 210, and 230 mm) were used along with 5 angular positions (O, 45,90,135,180 deg) g
41、iving 50 readings for each wheel. Corre- sponding to each probe position Pii(r, 0) is a constant flow area that is 4.02 x m2. The tests were performed at room temperature with a constant supply flow so flow properties were the same for each of the probe positions. The mass airflow rate to the probe,
42、 hP, was set to 1 .O8 x lo4 kg/s and was controlled to have a variation of less than 4.3 x 1 Od kg/s. Using the precision uncertainties for pressure and mass flow rate and its control, the uncertainty in the calculated diameter ratio as given by Equation 11 is U(d/d,) = *0.03 to 0.07 depending on th
43、e pressure drop. For each of these wheels we can calculate a mean value for Apo and the implied ratio of a standard deviation of pore diameter (o) divided by the mean pore diameter (Do) for the wheel. 1 1.8% t: 4.2% 7.71%f 1.10% (g) = standard deviation of (15) Wheel Matrix Parallel surface Honeycom
44、b Aluminum corrugated This pressure probe-determined ratio must be calibrated using the optical and micrometer data. These calculations for the honeycomb, aluminum corrugated, and paper corrugated wheels give (olD,) = 1.75%f0.44%, 2.52%f0.30%, and 7.71%fl. lo%, respectively. These pressure drop vari
45、ations and subsequent calcula- tion of pressure probe-determined hydraulic diameter ratios must now be corrected for bias using the corresponding opti- cal-micrometer results for each type of wheel. As noted previ- ously, the optical and micrometer determination of the dimensionless ratio of standar
46、d deviation to average pore (AP I APO) (El E3 85.9% 86.7% 96.6% 97.3% 98.2% 98.6% diameter will always exceed the corresponding pressure probe results. For the honeycomb and aluminum and paper corru- gated wheels these corrections were found to be 5.85%, 3.88%, and 4.09%, respectively. Adding these
47、bias corrections to the pressure probe-determined ratio of standard deviation in hydraulic diameter to mean hydraulic diameter (olD,) gives the corresponding optical results (o/D,). It is expected that similar correction factors could be obtained for other wheels, but each type and each manufacturin
48、g method will require an independent calibration of the pressure probe results. Paper corrugated SUMMARY AND CONCLUSIONS 94.5% 95.6% In this experimental study of regenerative wheels, four energy wheels were investigated for variation in flow channel pore sizes and the ratio of the standard deviatio
49、n of hydraulic diameter to mean hydraulic diameter for each wheel matrix. Two methods were used to measure these pore size variations, i.e., an optical and micrometer system for all four wheels and a pressure probe for three wheels. Both methods indicated a random variation in pore size, but the optical-micrometer system was regarded to be most accurate and without a signif- icant bias, while the more convenient to use pressure probe needs to be calibrated against the optical-micrometer results. A summary and comparison of the optical-micrometer and uncalibrate
