Você está na página 1de 14

CH-09-006 (RP-1319)

Measurements of Flat Oval Elbow Loss Coefficients

D. Kulkarni

S. Khaire

S. Idem, PhD
Member ASHRAE

This paper is based on findings resulting from ASHRAE Research Project RP-1319. ABSTRACT Pressure loss coefficients were determined for the following types of flat oval elbows having various aspect ratios: 5gore 90 easy bend, mitered 90 easy bend with and without vanes, and mitered 90 hard bend with and without vanes. The tests were performed in accordance with ANSI/ASHRAE Standard 120-1999. Data are presented graphically, and the loss coefficient for each fitting is tabulated. A linear correlation coefficient is calculated for each case. INTRODUCTION This paper presents results of an experimental program to determine loss coefficients for several flat oval elbows, i.e., 90 easy bend elbows, mitered 90 easy bend elbows with and without vanes, and mitered 90 hard bend elbows with and without vanes. A complete description of the test program is provided in Idem et al. (2008). In a hard bend flat oval elbow the fluid turns about an axis which is parallel to the minor axis of the fitting while in an easy bend elbow the fluid turns about an axis which is parallel to the major axis of the fitting. This research was motivated by the need to increase the population of flat oval fittings in the ASHRAE Duct Fitting Database (DFDB) (2006). Presently only straight flat oval duct and the 5-gore hard bend elbow are in the DFDB. EXPERIMENTAL PROGRAM Refer to Figure 1 for a sketch of the various elbows considered in this project. The dimensions of the flat oval and round elbows are listed in Tables 1 through 3. The test specimens were constructed using 20 gage galvanized steel. The ratio of major-to-minor dimensions ranged from 1.0 to 5.5. For the 5-gore easy bend elbow the dimensionless turning radius ratio R/a equaled 1.5. The elbow test setups shown in Figure 2 included a 30 hp centrifugal fan. A cylindrical nozzle chamber was used for flow measurement. A variable frequency drive (VFD) was used to control the fan speed, and hence the air flow through the system. Flow control screens were mounted upstream and downstream of the nozzle board to settle the flow. The system was blow through in nature. The nozzle board contained four long-radius spun aluminum flow nozzles having throat diameters of 51 mm (2 in.), 102 mm (4 in.), 152 mm (6 in.) and 203 mm (8 in.). The nozzles were mounted on a 25 mm (1 in.) thick plywood board. Various combinations of flow nozzles were selected to obtain a desired flow rate; unused nozzles were blocked by means of smooth vinyl balls. The nozzle pressure drop was measured by two piezometer rings located 38 mm (1.5 in.) on each side of the nozzle board, with both sides connected to a manometer. The nozzle chamber was constructed in accordance with ANSI/ASHRAE Standard 120-1999. Pressure taps constructed from 6.4 mm (1/4 in.) diameter copper tubing were soldered onto the ductwork upstream and downstream of the test section. Flexible tubing was used to construct piezometer rings at both measurement locations. The piezometer rings were connected to a single micromanometer using flexible tubing so as to measure the pressure drop across the test section. Static gage pressure was measured at each location by inserting tees into the pressure tubing. Pressure drop measurements over the test section and across the nozzle board were performed using liquid-filled

D. Kulkarni and S. Khaire are research assistants and S. Idem is a professor in the Department of Mechanical Engineering, Tennessee
Tech University, Cookeville, TN.

2009 ASHRAE

35

(a)

(b)

(c)

(d)

(e)

Figure 1 Flat oval elbows: (a) 5-gore 90 easy bend, (b) mitered 90 easy bend without vanes, (c) mitered 90 easy bend with vanes, (d) mitered 90 hard bend without vanes, and (e) mitered 90 hard bend with vanes.

36

ASHRAE Transactions

Table 1.
Fitting CF3-2 CF3-3 CF3-4 CF3-5 CF3-6 Geometry

Elbows Tested
Fitting Sizes See Table 2 See Table 2 See Table 2 See Table 2 See Table 2 See Table 3 for Number of Vanes See Table 3 for Number of Vanes Test Conditions R/a = 1.5

Elbow, Flat Oval, 5 Gore, 90, Easy Bend Elbow, Flat Oval, Mitered, Easy Bend, without Vanes Elbow, Flat Oval, Mitered, Hard Bend, without Vanes Elbow, Flat Oval, Mitered, Easy Bend, with Vanes Elbow, Flat Oval, Mitered, Hard Bend, with Vanes

Table 2.
Nominal Aspect Ratio 1.0 2.2 3.7 3.8 5.5

Nominal Elbow Dimensions


A a, mm mm (in. in.) 762 762 (30 30) 356 152 (14 6) 381 102 (15 4) 584 152 (23 6) 838 152 (33 6) 559 254 (22 10) 559 152 (22 6) 965 254 (38 10) 787 356 (31 14) 940 254 (37 10)

Table 3.

Number of Turning Vanes


Hard Bend with Vanes Major Axis, mm (in.) 356 (14) 381 (15) 559 (22) 584 (23) 787 (31) 838 (33) Number of Vanes 4 4 5 5 5 5

Easy Bend with Vanes Minor Axis, mm (in.) 102 (4) Number of Vanes 2

152 (6)

254 (10)

356 (14)

940 (37) 965 (38)

5 5

micromanometers having a measurement accuracy of 0.025 mm (0.001 in.). The gage pressure upstream and downstream of the test section was measured by means of inclined liquid-filled manometers having a readability of 0.25 mm (0.01 in.). Static pressure in the nozzle chamber was measured using an electronic manometer having a scale readability of 0.25 mm (0.01 in.). The air temperature in the nozzle chamber was measured using a mercury thermometer having a scale readability of 0.5C (1.0F). The dry- and wet-bulb temperatures of the ambient air were measured using an aspirated psychrometer with an accuracy of 0.5C (1.0F). The test section temperature was not measured directly, but was assumed to be the same as the temperature of the air inside the nozzle chamber. Ambient pressure was measured with a Fortin-type barometer, with an accuracy of 0.25 mm (0.01 in.) of mercury. All measurements of temperature and pressure in this project were in compliance with ANSI/ ASHRAE Standard 120-1999. Initially straight duct tests were performed so as to establish the tare pressure loss for subsequent evaluation of flat oval elbow loss coefficients. Darcy friction factors were determined by plotting the data on a Moody chart. The friction factor data were found to lie along a single relative roughness curve for each duct cross section that was tested, thereby validating the Colebrook equation model. Elbow pressure loss measurements were then performed by inserting the test fittings into the straight duct test apparaASHRAE Transactions

tus. The test apparatus and procedures complied with ASHRAE Standard 120-1999, except as noted below. In each instance the lengths of the upstream sections, the downstream section, and the tailpiece section are listed in Table 4 for each test configuration. In some instances the exit lengths for hard bend elbow tests did not conform to the values stipulated by ANSI/ASHRAE Standard 120-1999, due to the limited width of the laboratory. Those cases are marked by an asterisk in Table 4. The duct sections employed in most elbow tests were connected to the nozzle chamber by means of 305 mm (12 in.) diameter round to flat oval transitions, which did not conform precisely to ANSI/ASHRAE Standard 120-1999. This test setup is shown in Figure 2a. The exceptions were the pressure loss measurements performed on the 965 264 mm (38 10 in.) and 940 254 mm (37 10 in.) hard bend elbows with turning vanes. In those instances the elbow tests were conducted with a plenum chamber and bellmouth combination mounted upstream of the ductwork as depicted in Figure 2b. Likewise all 762 mm (30 in.) diameter elbow tests were performed using the plenum chamber and bellmouth mounted upstream of the test section. DATA ANALYSIS The loss coefficient is defined by Equation 1 as the ratio of the total pressure loss across an elbow to velocity pressure, where the cross sections (subscripts) are shown by Figure 2
37

(a)

(b) Figure 2 Elbow test apparatus: (a) with round to flat oval transition and (b) with plenum chamber.

38

ASHRAE Transactions

Table 4.
Duct Cross Section, mm mm (in. in.) 356 152 (14 6) 381 102 (15 4) 559 152 (22 6) 584 152 (23 6) 838 152 (33 6) 559 254 (22 10) 787 356* (31 14) 965 254* (38 10) 940 254* (37 10) 762 762* (30 30) Dh , mm (in.) 223.5 (8.8) 165.1 (6.5) 248.9 (9.8) 251.5 (9.9) 264.1 (10.4) 363.2 (14.3) 510.5 (20.1) 416.6 (16.4) 414.0 (16.3) 762.0 (30.0)

Elbow Test Setup Dimensions


LZ-1 , m (ft) 2.2 (7.2) 2.2 (7.3) 2.5 (8.3) 2.6 (8.4) 3.4 (11.0) 4.3 (14.0) 6.6 (21.5) 4.4 (14.3) 4.2 (13.7) 22.1 (72.5) L7-1 , m (ft) 0.3 (1.1) 0.2 (0.7) 0.4 (1.2) 0.4 (1.2) 0.3 (1.0) 0.6 (2.0) 0.8 (2.5) 0.5 (1.7) 0.7 (2.3) 1.1 (3.5) L2-8 , m (ft) 2.3 (7.7) 1.8 (5.8) 2.5 (8.2) 2.7 (8.7) 2.5 (8.3) 4.0 (13.2) 6.0 (19.6) 4.7 (15.3) 4.9 (16.2) 7.7 (25.3) Tail Duct Length, m (ft) 1.2 (4.0) 0.7 (2.3) 1.9 (6.3) 1.8 (5.8) 1.1 (3.7) 2.1 (6.8) 2.6 (8.4) 2.7 (8.7) 2.4 (7.8) 0.9 (2.8)

* Exit lengths for hard bend elbow tests did not conform to the values stipulated by ANSI/ASHRAE Standard 120-1999.

p t, 1-2 -. C = ---------------pv 7

(1)

pv f p f 1 ------- = -----------------------D h 1 1000 L pv f p f 1 -. ------- = -----------------L D h 1 12

(4 SI)

The velocity pressure at the test fitting is calculated based on the airflow and cross sectional area at section 7, thus Q 7 10002 2 1 - ---------------------- = 1 --V p v 7 = -2 A7 2 7 Q 7 A 7 - p v 7 = --------------- 1097
2

(4 IP)

(2 SI)

The least squares method was employed in order to obtain an overall loss coefficient for each elbow. Equation 1 can be written as p t, 1-2 = C p v 7. (5)

V7 2 - = ---------- 1097

(2 IP)

The total pressure loss across an equal area non-junction fitting (elbow) is calculated by Equation 3 p f - p t, 1-2 = p s, 7-8 ( L 7-1 + L 2-8 ) ------- L . (3)

The slope of the curve p t, 1-2 plotted against p v 7 can be interpreted as the zero length loss coefficient of the fitting, since a curve plotted through the data points is a straight-line if the loss coefficient is a constant. Hence, let y i = mx i + b where y i = p t, 1-2, x i = p v 7 and m = C. (8) (7) (6)

The terms L7-1 and L2-8 represent the separation distance between the upstream taps and the entrance plane of the elbow, and the exit plane of the elbow and the downstream pressure taps, respectively. In Equation 1, the pressure drop is due solely to dynamic losses in the duct fitting, since the friction pressure loss has been subtracted from the overall pressure drop across the fitting. The pressure friction loss per unit p f , as length is the duct tare pressure loss per unit length ------L calculated by Equation 4
ASHRAE Transactions

The intercept b can be forced to zero in the case of HVAC fittings, since ideally the pressure drop across the fitting should be zero when the velocity of the air through the fitting is zero. In that case it can readily be shown that
39


m =

xiyi (9)

i=1 ------------------. n 2 xi

Confidence intervals for the slope can be calculated under the assumption that the precision error in yi satisfies the normal distribution. The error in the slope of a least squares curve-fit (which in turn equals the uncertainty in the measured loss coefficient) is given by Beckwith et al. (1993) as sy x m = C = t ( a 2 ) n 2 --------S
xx

i=1

Measured data may contain both bias and precision (random) errors. Bias errors will either tend to shift the entire data set above or below the true line curve or change the slope. Precision errors will cause the data to scatter about the apparent line. The objective of curve-fitting is to average out the precision errors by calculating a curve that follows the apparent central tendency of the scattered data. The independent variable x or the dependent variable y may include both precision and bias errors. The least squares curve- fitting method implicitly assumes the precision error in y is much greater than that in x. Least squares curve-fitting cannot reduce bias error. The linear correlation coefficient is a measure of how the variance in y is accounted for by a linear curve-fit. It is interpreted as the ratio of the variation assumed by the fit to the actual measured variation in the data interpreted. Hence
2 explained variation -. r = --------------------------------------------total variation

(14)

where ta/2,n 2 is the students t-statistic with n 2 degrees of freedom (n is the number of points in the data set) and a = (1 c) is the level of significance (in the present work c = 95%).
n 2 2 1 s y x = ----------[ yi y ( xi ) ] n 2 i=1 1 --

(15)

where yi is the actual value at point i, and y(xi) was the value obtained by the least squares fit at point i. The quantity Sxx in Equation 14 is found from the expression S xx =
2 n

(10)

( xi x )
i=1

(16)

In general for a data set comprised of n variables the correlation coefficient can be calculated per Bethea et al. (1995) as

where x is the mean x value. RESULTS Loss coefficient measurements were performed for the flat oval elbows listed in Tables 1 through 3. Figures 3 through 12 depict the flat oval and round elbow loss coefficient data, plotted in terms of total pressure loss through the elbow as a function of velocity pressure. In each instance the slopes of the least squares curve-fit lines through the data correspond to the loss coefficients. Figures 3 through 12 are dimensionally correct for SI units only. Figures 10 and 11 include loss coefficient data for 965 254 mm (38 10 in.) and 940 254 mm (37 10 in.) hard bend elbows with vanes, respectively. In these cases the tare pressure loss was calculated using the relative roughness data obtained when a plenum chamber and bellmouth were mounted in the flow apparatus per Figure 2b. All other zerolength flat oval elbow loss coefficients determined in this project were based on the relative roughness data measured when an abrupt transition was used to connect the nozzle chamber to the ductwork, as shown in Figure 2a. The resulting loss coefficient data measured in this project are summarized in Table 5. Values of the linear correlation coefficient r obtained by means of Equation 11 are provided in Table 6. Similarly estimates of the error in the slopes of the least squares curves calculated per Equation 14 are included in Table 6.
ASHRAE Transactions

( xi x ) ( yi y )
i=1 r = --------------------------------------------------------------------n

( xi x ) ( yi y )
i=1 i=1

1 -2

(11)

where x and y are the mean values of x and y, respectively. They are defined by 1 x = -n

xi
i=1

(12)

and 1 y = -n
n

yi .
i=1

(13)

The linear correlation coefficient is a dimensionless quantity. In general 1 r + 1 , where a + sign indicates positive linear correlation and a sign implies negative linear correlation.
40

Figure 3 356 152 mm (14 6 in.) elbow loss coefficient data.

Figure 4 381 102 mm (15 4 in.) elbow loss coefficient data.

ASHRAE Transactions

41

Figure 5 559 254 mm (22 6 in.) elbow loss coefficient data.

Figure 6 559 254 mm (22 10 in.) elbow loss coefficient data.

42

ASHRAE Transactions

Figure 7 584 152 mm (23 6 in.) elbow loss coefficient data.

Figure 8 787 356 mm (31 14 in.) elbow loss coefficient data.

ASHRAE Transactions

43

Figure 9 838 152 mm (33 6 in.) elbow loss coefficient data.

Figure 10

965 254 mm (38 10 in.) elbow loss coefficient data.

44

ASHRAE Transactions

Figure 11

940 254 mm (37 10 in.) elbow loss coefficient data.

Figure 12

762 mm (30 in.) elbow loss coefficient data.

ASHRAE Transactions

45

Table 5.
A a, mm mm (in. in.) 356 152 (14 6) 381 102 (15 4) 559 152 (22 6) 584 152 (23 6) 838 152 (33 6) 559 254 (22 10) 787 356 (31 14) 965 254 (38 10) 940 254 (37 10) 5-Gore Easy Bend A/a 2.33 3.75 3.67 3.83 5.50 2.20 2.21 3.80 3.70 C 0.50 0.82 0.50 0.54 1.06 0.51 0.45 0.32 0.87

Pressure Loss Coefficients of Elbows

Mitered Hard Bend Mitered Hard Bend Mitered Easy Bend Mitered Easy Bend With Vanes No Vanes With Vanes No Vanes a/A C a/A C A/a C A/a C 0.43 0.27 0.27 0.26 0.18 0.45 0.45 0.26 0.27 5-Gore 1.07 1.79 1.45 1.42 1.68 0.91 0.90 1.10 1.23 0.43 0.27 0.27 0.26 0.18 0.45 0.45 0.26 0.27 1.68 2.44 2.11 1.51 2.40 1.66 1.45 2.61 1.91 2.33 3.75 3.67 3.83 5.50 2.20 2.21 3.80 3.70 1.03 1.93 1.03 1.22 1.53 0.89 0.69 0.51 1.17 2.33 3.75 3.67 3.83 5.50 2.20 2.21 3.80 3.70 1.45 2.17 1.40 1.81 2.00 1.65 1.50 1.26 1.73

762 762 (30 30)

A/a 1.00

C 0.44

MiteredWith Vanes A/a C 1.00 0.72

MiteredNo Vanes A/a C 1.00 1.49

Table 6.
Cross Section, mm mm (in. in.) 356 152 (14 6) 381 102 (15 4) 559 152 (22 6) 559 254 (22 10) 584 152 (23 6) 787 356 (31 14) 838 152 (33 6) 965 254 (38 10) 940 254 (37 10) 5-Gore Easy Bend r2 0.995 0.998 0.998 0.999 0.996 0.997 0.993 0.991 0.999

Flat Oval Elbow Loss Coefficient Correlation Analysis


Mitered Hard Bend Mitered Hard Bend Mitered Easy Bend Mitered Easy Bend With Vanes No Vanes With Vanes No Vanes r2 0.990 0.998 0.990 1.000 0.998 0.999 0.998 0.999 0.942

C
0.03 0.03 0.01 0.02 0.03 0.02 0.08 0.03 0.03
2

C
0.07 0.06 0.09 0.02 0.05 0.02 0.07 0.04 0.02

r2 0.993 0.998 0.996 0.999 0.999 0.997 0.999 0.996 0.999


2

C
0.12 0.08 0.08 0.05 0.04 0.08 0.08 0.16 0.08

r2 0.995 0.998 0.999 1.000 0.996 0.999 0.998 0.997 0.999

C
0.07 0.07 0.03 0.01 0.08 0.02 0.06 0.03 0.04
2

r2 0.993 0.997 0.997 1.000 0.998 0.999 0.995 0.998 0.999

C
0.10 0.09 0.06 0.02 0.07 0.07 0.13 0.06 0.02

762 762 (30 30)

5-Gore Easy Bend r 0.979

MiteredWith Vanes

MiteredNo Vanes r 0.994

C
0.05

r 0.978

C
0.10

C
0.10
ASHRAE Transactions

46

CONCLUSIONS Pressure loss tests were performed on 5-gore 90 easy bend flat oval elbows, mitered 90 easy bend flat oval elbows with and without vanes, and mitered 90 hard bend flat oval elbows with and without vanes. The measured total pressure loss across each elbow was plotted as a function of velocity pressure. A zero length loss coefficient was calculated by fitting a least squares curve to the pressure loss data. A loss coefficient table (loss coefficient as a function of aspect ratio and dimensionless hydraulic diameter) suitable for insertion into the DFDB was prepared for each fitting; these tables are available in Kulkarni et al. (2008). The 5-gore easy bend elbow consistently had the lowest loss coefficient of the various types of elbows tested, for every cross section. For hard bend and easy bend mitered elbows, those with turning vanes had a lower loss coefficient than the same size elbow without turning vanes. In general for each cross section the highest loss coefficients were associated with the hard bend mitered elbows without turning vanes. The quality of installation of turning vanes can impact the elbow loss coefficient, as improperly installed vanes can create more turbulence and pressure loss. In this project the vanes were mounted in the elbows strictly according to the specifications of the manufacturer. The aforementioned sudden transitions used to connect the nozzle chamber to the flat oval ducts per Figure 2a were chosen to minimize the overall extent of the test setups, in order to accommodate their lengths in the laboratory. It became apparent that test configurations having large major dimensions were particularly prone to flow-induced vibrations when such transitions were employed. Hence the test program was interrupted so as to increase the overall length of the laboratory by the removal of several walls and internal obstructions. This allowed insertion of a plenum chamber and bellmouth into the apparatus per Figure 2b, thereby bringing the setup into compliance with ANSI/ASHRAE Standard 120-1999. This was very successful, in that no random or excessive pressure fluctuations were apparent during these tests. In the future it is recommended that flat oval duct and elbow pressure loss tests be performed in strict accordance with ANSI/ASHRAE Standard 120-1999, particularly for flat oval geometries with large major spans. It is likewise suggested that flat oval elbows having additional aspect ratios and hydraulic diameters (including round elbows having an aspect ratio of unity) be tested, in order to augment the database. ACKNOWLEDGMENTS The work reported in this paper is the result of cooperative research between ASHRAE (RP-1319) and Tennessee Tech University. The project was sponsored by TC 5.2, Duct Design, and their technical assistance is gratefully acknowledged. The authors are deeply indebted to Brad Thomas, Chair of TC 5.2 and Vice-President of Hamlin Sheet Metal Company, Inc., who donated the ducts, elbows, and bellmouths used in this study.
ASHRAE Transactions

NOMENCLATURE A a b C c Dh f L m n pv R r Sxx sy/x V xi, yi x ,y p f p s p t C = major duct dimension, mm (in.) or cross-sectional duct area, m2 (ft2) = minor duct dimension, mm (in.) = intercept of fitted line to data = loss coefficient, dimensionless = confidence interval, % = hydraulic diameter, mm (in.) = friction factor, dimensionless = length, m (ft) = slope of fitted line to data = number of data points = velocity pressure, Pa (in. wg) = elbow turning radius, mm (in.) = linear correlation coefficient, dimensionless = total squared variation = standard error of y-data about the curve fit = average velocity, m/s (ft/min) = variables = mean values = frictional pressure losses, Pa (in. wg) = static pressure differential, Pa (in. wg) = total pressure differential, Pa (in. wg) = loss coefficient uncertainty, dimensionless = air density, kg/m3 (lbm/ft3)

Subscripts 1 2 7 8 = = = = plane 1 plane 2 plane 7 plane 8

Acronyms EB HB = easy bend = hard bend

REFERENCES ANSI/ASHRAE Standard 120-1999. Method of Testing to Determine Flow Resistance of HVAC Ducts and Fittings. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. ASHRAE. 2006. Duct Fitting Database. Version 4.0.3. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. Beckwith, T.J., R.D. Marangoni, J.H. Lienhard V. 1993. Mechanical Measurements, 5th Edition. New York: Addison-Wesley Publishing Company, Inc. Bethea, R.M., B.S. Duran, and T.L. Boullion. 1995. Statistical Methods for Engineers and Scientists, 3rd Edition. New York: Marcel Dekker, Inc. Idem, S., D. Kulkarni, and S. Khaire. 2008. Laboratory Testing of Duct Fittings to Determine Loss Coefficients. Final Report, ASHRAE RP-1319. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. Kulkarni, D., and Khaire, S. and Idem, S. 2008. Influence of Aspect Ratio and Hydraulic Diameter on Flat Oval Elbow Loss Coefficients. ASHRAE Transactions, In Press.
47

Você também pode gostar