1、NA-04-1 -1 Viscosity Measurements and Model Comparisons for the Refrigerant Blends R-410A and R-507A Arno Laesecke, Dr.-lng. ABSTRACT Wide-ranging viscosity measurements of the blends R- 410A (0.5 R-32 i- 0.5 R-125 by mass) andR-507A (0.5 R-143a i- 0.5 R-i25 by mass) were carried out in a torsional
2、cvstal viscometer at two subcritical and three supercritical isotherms between 300 and 420 K with pressures up to 82 MPa. Consid- erable conductances were observed in the blend containing difluoromethane (R-32). Therefore, reference viscosity measurements of the blend R-41 OA were carried out at sat
3、u- rated-liquid conditions between 240 K and 31 O K with a sealed gravitational capillary viscometer; because in that instrument the sample is not exposed to an electricfield. The measurement results are compared with values estimated via the extended- co responding-states model in NISTStandard Refe
4、rence Data- base 23 (REFPRO8 version 7.0) und with literature data. For both blends, the results obtained with the torsional crystal viscometer agree with the predicted viscosities within the esti- matedexperimentaluncertang ofthe instrument at low sample conductances. The deviations increase system
5、atically with the sample conductance. The data for R-410A obtained with the capillary viscometer agree with the estirnatedviscosities within their experimental uncertain. Literature data deviations rangefrom -18% to 5% for this blendandfrom -9% to 14% for R-507A, while the R-507A-measurements of thi
6、s work agree with the model viscosities within *4%. INTRODUCTION Significant energy penalties have been measured in HVAC systems with next-generation alternative refrigerant blends operating at condenser temperatures that are closer to the critical temperatures of the blends. To understand these ene
7、rgy penalties, working fluids have to be characterized not only at subcritical saturated-vapor and liquid states but also at supercritical conditions. Literature viscosity data for such blends and working conditions are very sparse. Accordingly narrow is the basis on which predictive viscosity model
8、s for such mixtures have been developed. Besides, substantial errors have been identified frequently in literature viscosity data for hydrofluorocarbon systems. Therefore, additional viscosity measurements are needed to broaden the base of validated and reliable data for these important fluids. Meas
9、ur- ing the blend R-410A in two different viscometers presented an opportunity to discern the susceptibility of this fluid to an applied electric field and to quantify electroviscous effects in the results of the torsional crystal viscometer. The report is organized as follows. In an experimental se
10、ction, the preparation of the mixture samples is described, and the torsional crystal viscometer and data analysis are explained. Results and comparisons with the extended-corre- sponding-states viscosiy model and literature data are presented first for the blend R-507A, then for R-410A, because the
11、 measurements of the latter were more complex. EXPERIMENTAL SECTION Mixture Preparation The test mixtures R-410A (R-32/125) and R-507A (R- 125/143a) were prepared gravimetrically in aluminum cylin- ders with avolume of 15.8 dm” from high-purity components. R-32 (difluoromethane, CH,F,) and R-125 (pe
12、ntafluoroet- hane, C,HF,) were “electronic grade” materials. The manu- facturers specification stated a minimum purity of 99.99% with a water content less than 5 ppm. In-house gas chromato- graphic analysis confirmed the high purity apart from a small Arno Laesecke is a chemical engineer in the Phys
13、ical and Chemical Properties Division of the Chemical Science and Technology Laboratory, National Institute of Standards and Technology, Boulder, Colo. 02004 ASHRAE. 503 Component 1 Component 2 X1 x2 W1 R-4 1 OA R-32 R-125 0.69767 0.30233 0.50007 R-507A R-143a R-125 0.58832 0.41 168 0.50016 amount o
14、f air in the vapor phase. R-143a (l,l,l-trifluoroet- hane, CH,-CF,) was refrigerant-grade material. Our analysis indicated a purity of 99.93%. Air and any other light inorgan- ics, which may have been present in the components, were removed by repeatedly freezing the sample in liquid nitrogen, evacu
15、ating the vapor space, and thawing. The target compo- sitions for both blends were 0.50 by mass fraction. The actual compositions of the test samples are given in Table 1. w2 0.49993 0.49984 Torsional Crystal Viscometer The measuring element in the torsional crystal viscometer is a piezoelectric qua
16、rtz crystal of cylindrical shape. The crys- tal is mounted in a transducer and driven into a torsional oscil- lation by an ac voltage applied to two of four surrounding brass electrodes. The frequency of the voltage is varied through the mechanical resonance of the crystal. The resonant frequency an
17、d the bandwidth of the resonance depend on the damping by the fluid surrounding the crystal. Absolute measurements of fluids can be performed by comparing the damping to that in vacuo. The technique and details of the instrument were described by Diller and Frederick (1989) and by Hafer and Laesecke
18、 (2003). The working equation for absolute viscosity measure- ments with the torsional crystal viscometer in the frequency domain is where q denotes the viscosity of the fluid, p its density, m the mass and S the surface area of the crysta1,JPC the resonant frequency, and Afthe bandwidth of the reso
19、nance. Subscript O indicates vacuum conditions. Since the technique measures the product (q x p), the density of the fluid must be obtained separately in order to determine its viscosity and the uncer- tainty of the density propagates into the uncertainty of the viscosity measurement. In this work,
20、densities of the refriger- ant blends were calculated for the measured compositions, pressures, and temperatures with the mixture models in NIST Standard Reference Database 23, REFPROP (Lemmon et al. 2002). Their estimated uncertainty is kO.l%. The viscosity measurements were carried out with the cy
21、lindrical quartz crystal “D (diameter d = 3.046 mm, length L = 50.076 mm, mass m = 0.9662 g) in the transducer of the viscometer. The resonance of the crystal in vucuo depends on the applied voltage to drive the torsional crystal vibration and on temperature. The change of the resonant frequency fo*
22、 and of the resonance bandwidth R-507A was measured first and R-410A second. The measurements consist of frequency scans of the transducer resonance in the sample fluid at every pressure of each isotherm and at several drive voltages. Resolving the dependence of the resonance on the drive voltage al
23、lows extrapolation to the resonance at zero drive voltage, which is consistent with the definition of Newtonian viscosity as a property characterizing momentum transport in the limit of 504 ASHRAE Transactions: Symposia G Ips q ImPa-s 0.35 0.30 I e300 K I I l 0.6 0.5 0.4 0.3 0.2 0.1 0.0 O 200 400 60
24、0 800 1000 1200 1400 Figure I Conductances G at 39 kHz of the mixture R-507A (0.5 R-143a + 0.5 R-125 by mass) as measured in the torsional clystal viscometer. zero shear. Due to the resolution limits of the impedance analyzer, the majority of the measurements were limited to drive voltages from O. 1
25、 V to 1.1 V. Lower voltages, down to 1 O mV, could be applied only in a few cases of dilute gases. For reasons of consistency, all the results reported in this work were derived fi-om resonance scans with drive voltages of O. 1 V. The uncertainty of measurements with the torsional crystal viscometer
26、 typically does not exceed *2%. Measurements of fluids with a background conductance G at 39 kHz of more than 1 pS may be influenced by electroviscous effects because the sample fluid is affected by the applied drive voltage. The experimental viscosities measured with the torsional crystal viscomete
27、r were compared with viscosities calculated with the extended corresponding states model in REFPROP (Klein et al. 1997; Lemon et al. 2002) using the recently developed correlation for the viscosity of R- 134a as the refer- ence fluid (Huber et al. 2003). They were also compared with literature data.
28、 In the following, the results will be presented for each blend, and their reliability will be discussed. RESULTS AND DISCUSSION Results for R-507A The impedance analyzer in the viscometer measures the conductance G (in units of siemens or reciprocal ohm) of an 0.25 0.20 0.15 0.10 0.05 O o 200 400 6
29、00 ao0 1000 1200 i400 p I kgsm Figure 2 Density dependence of experimental viscosities of the mixture ofR-507A (0.5 R-143a i- 0.5 R-125 by mass) as measured in the torsional crystal viscometer. electrical circuit that includes the transducer with the crystal and the surrounding fluid in the cell. Th
30、e conductance values far away from the mechanical resonant frequency of the vibrating crystal provide valuable information about the sample fluid and its polarity because measurements of fluids with elevated background conductance (G 1 pS) may be influenced by electroviscous effects when the sample
31、fluid becomes susceptible to the applied drive voltage. HFC refrig- erants are rather polar compounds. Figure 1 shows the measured conductances G at 39 kHz of the blend R-507A as a function of density. Any increase of the conductance beyond that of the instrument circuitry when the cell is evacuated
32、 indicates the presence of charge carriers in the sample. No increase of the conductances of R-507A occurs up to a density of about 500 kg.m3. At higher densities, the conductances increase with both density and temperature. While the increase is monotonic at 360 K and above, a flat- tening is obser
33、ved on the isotherms at 300 K and 340 K begin- ning at about 1100 kg.m-3. The highest conductance of G = 0.62 pS was measured at the highest pressure of the 390 K- isotherm. The scatter of the conductances at 360 K was caused by intermittent problems with an electrical connection in the instrument c
34、ircuitry. The viscosities for the blend R-507A measured in the torsional crystal viscometer are tabulated in Table 2. Their density dependence is displayed in Figure 2. Due to the measurements at supercritical temperatures, the entire density range from 0.09 kg.m“ to 1300 kg.m“ is covered and a larg
35、e data gap existing in the literature has been filled. Figure 2 shows that the viscosity data aggregate along a single line with a barely discernible temperature dependence. This depen- ASHRAE Transactions: Symposia 505 Table 2 . Experimental Viscosities of the Mixture R-507A (0.5 R-143a + 0.5 R-125
36、 by mass) as Measured in the Torsional Crystal Viscometer The contents of each line represents the average of four measurements. The precision of the measurements is indicated by the relative standard deviation of the viscosity aver- aging, s,/q. Calculated density values were obtained from REFPROP
37、version 7.0. P T P Il sqh MPa measured 66.852 54.147 41.734 3 1.904 23.731 16.841 1 1.202 6.595 3.575 K kgm” mPa-s */O measured calculated measured measured 301.29 1299.9 0.2924 0.73 300.79 1273.7 0.2661 0.82 300.52 1243.1 0.2395 0.32 300.33 1213.8 0.2 156 0.29 300.15 1184.7 O. 1968 0.63 299.99 1 15
38、4.7 0.1763 0.30 299.85 1 124.4 O. 1599 0.32 299.74 1092.9 O. 1457 0.26 299.66 1066.6 O. 1344 0.56 67.754 52.722 44.042 38.680 3 1.652 25.845 20.764 16.782 13.648 11.013 8.961 7.292 6.043 5.204 4.437 3.986 3.686 3.485 3.059 2.767 2.394 2.053 1.598 1 .O02 0.2657 340.65 340.65 340.66 340.65 340.64 340.
39、66 340.65 340.65 340.66 340.66 340.67 340.66 340.67 340.65 340.66 340.66 340.65 340.66 340.65 340.66 340.66 340.68 340.66 340.67 340.66 1229.9 1190.4 1162.8 1143.3 1113.8 1084.5 1053.5 1023.7 994.87 964.72 934.99 903.87 872.99 845.52 809.83 778.36 745.40 702.86 188.16 152.43 119.14 94.81 1 68.140 39
40、.186 9.53 14 0.2239 O. 1960 0.1804 0.1688 O. 1537 0.1414 0.1297 0.1187 O. 1097 0.1013 0.09395 0.08628 0.07944 0.07488 0.06844 0.06222 0.05733 0.052 13 0.01693 0.01626 0.0 1526 0.01457 0.01462 0.01422 0.0 1401 0.34 0.24 0.42 0.48 0.16 0.83 0.10 0.39 0.13 0.20 0.10 0.17 0.16 0.27 0.14 0.18 0.19 0.14 O
41、. 14 0.85 1.75 1.12 0.57 0.61 0.27 506 ASHRAE Transactions: Symposia Table 2 (continued). Experimental Viscosities of the Mixture R-507A (0.5 R-143a + 0.5 R-125 by mass) as Measured in the Torsional Crystal Viscometer The contents of each line represents the average of four measurements. The precisi
42、on of the measurements is indicated by the relative standard deviation of the viscosity aver- aging, s Laesecke et al. 1999). The data analysis of the R-410A measurements was based on the measured bulk density = 343.4 kgm-3 of the sample in the viscometer. The contribution of the uncertainty of the
43、bulk density to the exper- imental uncertainty of the viscosity measurements was assessed by comparing the calculated sample properties at the measured bulk density with those at bulk densities 10% below and above that value. The resulting variations in the saturated liquid and vapor density ranged
44、from -0.014% to 0.017% and from -0.055% to 0.067%, respectively. Thus, the data analysis is rather insensitive to the uncertainty of the bulk density. The expanded uncertainty of the present measurements is estimated at i2.4% (coverage factor of 2), the same level as in our previ- ous measurements o
45、f the similar system R-32 + R134a (Laesecke and Hafer 1998). The results from the capillary viscometer are presented in Table 4 and included in Figure 6. The temperature dependence of the viscosity of R-41 OA is more noticeable than in the results for R-507A. A magnified 512 ASHRAE Transactions: Sym
46、posia Table 3 . Experimental Viscosities of the Mixture R40A (0.5 R-32 + 0.5 R-125 by mass) as Measured in the Torsional Crystal Viscometer. The contents of each line represents the average of four measurements. The precision of the measurements is indicated by the relative standard deviation of the
47、 viscosity aver- aging, Snh. Calculated density values were obtained from REFPROP version 7.0. Values in italie not to be used in model development (see text). P MPa measured 82.477 75. O 79 59.368 48.136 40. I72 32.85 7 26.608 20.770 15.808 11.463 7.820 4.797 2.530 67.71 6 56.626 49.354 42.629 36.2
48、30 32.250 27.851 23.520 20.338 17.696 15.043 13.208 1 1.298 9.774 8.142 7.200 6.586 5.930 5.426 4.398 4.293 T K measured 3 04.72 304.79 304.80 304.84 304.87 304.87 304.90 304.85 304.78 304.69 304.56 304.43 304.58 340.38 340.38 340.38 340.38 340.38 340.3 7 340.3 7 340.36 340.38 340.38 340.37 340.36 3
49、40.38 340.40 340.36 340.38 340.37 340.37 340.37 340.38 340.37 1309.7 1296.5 1265.1 1238.8 1217.5 1195.3 1173.6 Il 50.5 1127.7 1104.2 1080.9 1057.5 1034.4 1208. I 1179.3 1157.7 1135.1 1110.6 1093.4 1071.9 1047.6 1026.8 1007.0 983.96 965.38 942.54 920.65 891.72 870.20 853.30 830.95 808.46 250.68 226.82 O. 2653 0.2552 0.2280 O. 21 04 O. 1986 O. 1842 O. 1723 O. 1623 0.1512 0.1419 O. 1325 0.1241 0.1164 0.1987 o. I799 O. 1681 O. 1562 O. I453 0.1384 O. 1303 0.1217 0.1147 O. 1086 0.1026 0.09701 0.09143 0.08652 0.08004 0.07573 0.07263 0.06874 0.06573 0.01 8 72 0.017992 YO measured -
