ICR0026

FIN EFFICIENCY CALCULATION IN ENHANCED FIN-AND-TUBE HEAT EXCHANGERS IN DRY CONDITIONS

Thomas PERROTIN, Denis CLODIC Ecole des Mines de Paris, Center for Energy Studies, 60, boulevard Saint-Michel, F-75272 Paris Cedex 06 Phone: +33 1 40 51 92 49, Fax: +33 1 46 34 24 91, email: thomas.perrotin@ensmp.fr

ABSTRACT
Fin efficiency calculation is of the greatest importance in refrigerant-to-air heat exchanger engineering, for the evaluation of the finned surface performance or for the determination of the air-side heat transfer coefficient from experimental data. High efficiency heat exchangers use enhanced fin geometry (louvered and slit fins) for which the fin efficiency could be overestimated by usual formulations and more precisely equivalent circular fin and conventional 1-D sector methods. Because the slits (or louvers) alter the conduction path through the fin, the assumption of radial heat flow pattern is no more valid, and the actual fin efficiency could only be determined by numerically solving the multi-dimensional heat conduction equation. SimTherm, a software developed by the Center for Energy Studies, is used for this purpose (2-D) and results are compared to commonly used fin efficiency formulations.

INTRODUCTION
Fin-and-tube heat exchangers are widely used in several domains such as heating, ventilating, refrigeration and air conditioning systems. In practical application of air-to-refrigerant heat exchangers, the dominant resistance is on the air-side and improving the accuracy of the analysis of the air-side heat transfer is required by the growing demand of high performance heat transfer surfaces. The fin performance is commonly expressed in terms of heat transfer coefficient and fin efficiency, which is defined as the ratio of the actual fin heat transfer rate to the heat transfer rate that would exist if all the fin surface was at the base temperature. This case is the one providing the maximum heat transfer rate because this corresponds to the maximum driving potential (temperature difference) for the convection heat transfer. Many experimental studies available in the open literature have been performed in order to characterize the air-side heat transfer performance of several type of fins used in finned tube heat exchangers [1] [2] [3], and establish correlations which are used for design, rating and modeling of heat exchangers. In order to obtain the heat transfer coefficient, it is necessary to determine the fin efficiency [4]. What is observed in nearly all published papers is that, whatever the fin type (plain, louvered, slit), the fin efficiency calculation is always performed by analytical methods derived from circular fin analysis. When the heat transfer coefficient h is considered separately from its corresponding fin efficiency calculation (used for h measurement), error could be generated. If h is always associated to the fin efficiency calculation that served for h measurement, there is no possible error. The analytical circular fin analysis involves a number of assumptions which need to be addressed. These assumptions, known as ideal fin assumptions (attributed to Gardner [5]), are: 1. 1-D radial conduction, 2. steady state conditions, 3. radiative heat transfer negligible, 4. constant fin conductivity, 5. constant heat transfer coefficient over the entire fin, 6. the fin base temperature is assumed to be constant, 7. the thermal contact resistance between the prime surface and the fin is negligible, 8. the surrounding fluid is assumed at constant temperature. In the present study, the commonly used analytical methods for fin efficiency calculation in finned tube heat exchangers are reviewed and compared. Among the ideal fin assumptions, the first one should be carefully considered because the actual fin geometry used in finned tube heat exchanger differs significantly from the plain circular fin shape. In particular, for enhanced fin designs with louvers or slits, the fin shape alters the conduction path within the fin. 2-D numerical models are used in order to quantify the deviation generated by the 1-D assumption, depending on the fin geometry and type. International Congress of Refrigeration 2003, Washington, D.C. 1

dqconv dR

Pl 2 XL = Pl 2XD 2 XT = P t 2 XT = Pt

R Adiabatic fin side 1-D radial heat flow pattern

Fin base at Tb

Figure 1: Control volume in an angular fin of rectangular profile

Figure 2: Unit cells for inline and staggered tube layouts with plain plate fins

1 ANALYTICAL METHODS
Fin efficiency equations for dry plain circular fins under the aforementioned assumptions are reported in many handbooks [6] [7]. The analytical solution for a circular fin, which is the same as for an angular sector of circular fin as presented in figure 1, with adiabatic fin tip is given by equation (1), where In and Kn are the modified Bessel functions of first and second kind.
f =
K1 ( m r ) I 1 ( m r f ) K1 ( m r f ) I1 ( m r ) 2r m (r f2 r 2 ) K1 ( m r f ) I 0 ( m r ) + K 0 ( m r ) I 1 ( m r f )

(1)

Several studies have been performed in order to simplify this circular fin efficiency formulation by avoiding the use of modified Bessel functions. Among all the approximations, the Schmidt approximation [8] is the most widely used one. Hong and Webb [9] propose to slightly modifiy the Schmidt equation in order to obtain better accuracy (eqs. 2 and 3). In the present study, it is proposed to use a modified parameter (eq. 4) in equation (2). With this formula, the error between the analytical solution (eq. 1) and the approximation does not exceed 2% over the practical range of conditions rf/r 6 and m(rf r) 2.5.
f =
tanh(m r ) ; cos(m r ) m = mr

2h

(2)

f f

rf rf 1 + 0.35 ln 1 r r

(3)

1 rf 1.5 rf m (r f r ) 12 r m = r 1 1 + 0.3 + 2.5

rf 0.26 r

0.3

rf 0.3 ln r

(4)

Fin and tube heat exchangers are generally composed of continuous plate fins. The fins are metal sheets pierced through by the tube bank. The tube lay-out is in inline or staggered configuration (fig. 2), with a clear advantage for the staggered lay-out. In order to express the fin efficiency of such continuous plate fins, the fin is divided in unit cells. Considering that all the tubes are at the same temperature, the adiabatic zones of the fins determine the unit cells, as presented in figure 2. The considered fin shape is rectangular for the inline configuration and hexagonal for the staggered lay-out. Two methods are used in order to calculate the efficiency of these rectangular or hexagonal fins from the circular fin efficiency with adiabatic fin tip condition. The most accurate method is the sector method. Nevertheless, the equivalent circular fin method is simpler and is more widely used.

1.1 Equivalent circular fin methods

Gardner [5] and Schmidt [10] have shown that in the case of rectangular and hexagonal fins, the fin efficiency could be treated as for a circular fin, by considering an equivalent circular fin radius. For the calculation of the equivalent circular radius, two approaches are possible. The first one consists in considering a circular fin having the same International Congress of Refrigeration 2003, Washington, D.C. 2

surface area as the rectangular or hexagonal fin. The other method is the Schmidt method in which correlations are developed in order to find an equivalent circular fin having the same fin efficiency as the rectangular fin (eq. 5) or the hexagonal fin (eq. 6).
r f ,eq r r f ,eq r XT r = 1.28 XT r XL 0.2 XT

(5)

= 1.27

P2 XD ; 2 2 2 0.3 2 X D = Pl + t = 4 X L + X T 4 XT

(6)

1.2 Sector method

The sector method could be characterized as a semi-analytical method. The plain fin surface is divided in several circular sectors generated from the tube center and fitting the fin geometry profile (figures 3 and 4). The inner radius of each sector is equal to the tube radius while the outer radius is equal to the tube center to fin tip distance, corresponding to the considered sector. Doing so, it is possible to approximate every kind of fin profile. Considering that the conductive thermal flux through each angular sector is purely radial, the rectangular or hexagonal fin efficiency is the surface weighted average of the efficiency of each sector (eq. 7). With the radial flux assumption, the lateral faces of each sector are adiabatic. The sector base is at constant temperature and the sector tip is considered adiabatic. Consequently, the sector efficiency is analytically evaluated from the circular fin efficiency formulas, equation (1) for the exact solution with Bessel functions, or other approximated equation (Schmidt, Hong and Webb, eqs. 2 to 4).

f =

A
i =1 n i

(7)

A
i =1

25 20 15

20

Pl/2=XL XL
15

10 10

XT
5 0 -45 -35 -25 -15 -5 -5 -10 5 15 25 35 45 -35 -25 -15 -5

Pt/2=XT
5

XD/2

0 5 -5 15 25 35

-10 -15 -20 -25 -15

-20

Figure 3: Sector method applied to square fin XL/XT = 1 ; XT/r = 2

Figure 4: Sector method applied to equilateral hexagonal fin PL/PT = 0.866 ; PT/2r = 2

1.3 Comparison of the analytical methods

Figures 5 and 6 present, for square, rectangular and hexagonal profiles, the comparison between the fin efficiency calculated by the sector method and the fin efficiency evaluated by the equivalent circular fin having the same surface area. It can be seen that for the rectangular fin profile, the equivalent circular method overestimates the fin efficiency compared to the sector method. For the square fin profile, aspect ratio of 1, the circular method shows very good agreement with the sector method, within 2.5% over the considered range of m.XT parameter. For rectangular fin, when the aspect ratio of the fin is important, the equivalent circular fin method shows unacceptable error. For a XL/XT ratio of 3, the error is up to 30%. The error between the equivalent circular method and the sector method is significantly reduced (~10%) for rectangular fins with poor fin efficiency, i.e. for high values of m.XT or high XT/r ratios.

International Congress of Refrigeration 2003, Washington, D.C. 3

For the equilateral hexagonal fin profile (figure 6), the equivalent circular fin of same area underestimates the fin efficiency compared to the one calculated by the sector method. The equivalent circular method shows unacceptable error up to 10% over the considered range of m.XT parameter. Figures 7 and 8 present, for square, rectangular and hexagonal profiles, the comparison between the fin efficiency calculated by the sector method and the fin efficiency evaluated by the Schmidt equivalent circular fin method (with eqs. 5 or 6). For the square fin profile and for the equilateral hexagonal fin profile, the Schmidt method shows very good agreement with the fin efficiency calculated by the sector method. Both for square and equilateral hexagonal profiles, the error does not exceed 2.2% over the considered range of m.XT parameter. Nevertheless, when the fins are no longer equilateral, the error between the Schmidt equivalent circular fin method and the sector method increases. For an aspect ratio of 2, the error is within 10% and for an aspect ratio of 3, the error increases up to 20%. The circular fin of the same area and the Schmidt circular fin methods are generally comparable but advantage is given to Schmidt method which significantly more accurate for hexagonal fin profiles. Both of these methods show unacceptable error for high aspect ratio fin profiles.
1,4

Rectangular fin XL / XT = 3

1,1

1,3 f,circular / f,sector

XT/r = 2 XT/r = 3 XT/r = 4

f,circular / f,sector 1

1,2

1,1

Square fin XL / XT = 1

0,9

XT/r = 2 XT/r = 3 XT/r = 4

Equilateral hexagonal fin XL / XT = 0.866

0,9 0 0,5 1 m XT 1,5 2 2,5

0,8 0 0,5 1 m XT 1,5 2 2,5

Figure 5: Ratio of rectangular fin efficiencies calculated by eq. circular method (same area) and sector method
1,4 1,3 f,circular / f,sector 1,2 1,1 1 0,9 0,8 0 0,5 1 m XT 1,5 2 2,5

Figure 6: Ratio of hexagonal fin efficiencies calculated by eq. circular method (same area) and sector method
1,4 1,3 f,circular / f,sector 1,2 1,1 1 0,9 0,8 0 0,5 1 m XT 1,5 2 2,5

Rectangular fin XL / XT = 3

XT/r = 2 XT/r = 3 XT/r = 4

XT/r = 2 XT/r = 3 XT/r = 4 Hexagonal fin XL / XT = 2

Square fin XL / XT = 1

Equilateral hexagonal fin XL / XT = 0.866

Figure 7: Ratio of rectangular fin efficiencies calculated by eq. circular method (Schmidt) and sector method

Figure 8: Ratio of hexagonal fin efficiencies calculated by eq. circular method (Schmidt) and sector method

2 RADIAL HEAT FLOW ASSUMPTION

2.1 Fin shape and fin aspect ratio
It is obvious that when the fin aspect ratio is near 1, the actual conductive path is mainly radially directed, and the 1D assumption is reasonable, but for rectangular fins or for non-symmetrical fins, this assumption becomes non-valid. In order to quantify the validity of radial heat flow pattern assumption for rectangular and hexagonal fins of high aspect ratio, the 2-D heat equation is solved numerically by finite difference method with SimTherm, a software developed by the Center for Energy Studies. These 2-D calculations are performed under the ideal fin assumptions, with the ideal 1-D assumption excluded. In particular, the heat transfer coefficient is assumed constant over the entire fin surface.

International Congress of Refrigeration 2003, Washington, D.C. 4

Figure 9 presents the temperature distribution for different calculation methods and different fin shapes. Figure 9a) presents the temperature distribution obtained by the sector method (1-D assumption) in the case of a rectangular fin with XL/XT = 2. Comparing this figure to figure 9b) shows that the temperature distribution with sector method differs significantly from the 2-D computation. The 2-D model taking into account for the transversal conduction shows more homogenous temperature distribution and the temperature gradient is smaller than in the case of the sector method. Consequently it is expected that the actual fin efficiency is higher than the one calculated by the sector method. Figures 10 and 11 present the ratio of the fin efficiencies calculated by numerical 2-D method and sector method, for rectangular and hexagonal fins. It is observed that, in the case of rectangular fins, the fin efficiency calculated by the sector method is underestimated compared to the numerical 2-D solution. As expected, the difference between the 2-D method and the sector method increases when the fin aspect ratio is increased. When the aspect ratio of the fin XL/XT is 2, the sector method underestimates the fin efficiency up to 5.5% in the pratical range of m.XT parameter, while for square fin, the error does not exceed 1%. For hexagonal plain fins (figure 11), the sector method and the 2-D method are in better agreement but the same trends are observed. When the aspect ratio is near 1 as presented in figure 9c), the temperature distribution obtained by 2-D numerical simulation is very close to the circular fin temperature distribution. Consequently, it is observed in figure 11 that the sector method is in very good agreement with the 2-D calculation. When the ratio PL/PT is under 1 (i.e. when the longitudinal tube pitch is smaller than the transversal tube pitch, which is often the case in practice), it is observed that the sector method slightly overestimates the fin efficiency. When the ratio PL/PT is increased, the difference between the 2-D solution and the sector method is increased. The sector method underestimates the fin efficiency up to 4.5% for a ratio PL/PT=2. a) b) c)

Figure 9: Temperature distribution (K) within rectangular and hexagonal fins m.XT = 0.582 ; h = 40W/m2.K ; f = 170W/m.K ; f = 0.1mm ; Ta = 27C ; Tbase = 54C a) Rectangular fin ; XL/XT = 2 ; XT/r = 2 ; Sector method b) Rectangular fin ; XL/XT = 2 ; XT/r = 2 ; SimTherm computation c) Hexagonal fin ; PL/PT = 1 ; PT/2r = 2 ; SimTherm computation
1,06
1,06

1,04 numeric / sector

numeric/sector

1,04

1,02

1,02

1 XL/XT = 1 XL/XT = 2 XL/XT = 1.75 XL/XT = 1.5 XL/XT = 1.25

0,98

0,98

PL/PT=0.866 PL/PT=2 PL/PT=1.5

0 0,3 0,6 0,9 m XT 1,2 1,5 1,8

0,96 0 0,3 0,6 0,9 m XT 1,2 1,5 1,8

0,96

Figure 10: Ratio of rectangular fin efficiencies calculated by numerical 2D model and sector method

Figure 11: Ratio of hexagonal fin efficiencies calculated by numerical 2D model and sector method

International Congress of Refrigeration 2003, Washington, D.C. 5

1,06

1,06

1,04 numeric / circular schmidt

numeric /circular schmidt

1,04

1,02

1,02

1 XL/XT = 1 XL/XT = 1.5 XL/XT = 2 0 0,3 0,6 0,9 m XT 1,2 1,5 1,8

0,98

0,98

PL/PT=0.866 PL/PT=2 PL/PT=1.5

0 0,3 0,6 0,9 m XT 1,2 1,5 1,8

0,96

0,96

Figure 12: Ratio of rectangular fin efficiencies calculated by 2D model and equiv. circular fin method

Figure 13: Ratio of hexagonal fin efficiencies calculated by 2D model and equiv. circular fin method

Since the equivalent circular fin method (Schmidt method) is more widely used than the sector method, it is important to compare the numerical 2-D solution to the equivalent circular fin method. Figures 12 and 13 show the fin efficiency ratio for rectangular and hexagonal fins, respectively. For rectangular fins (fig. 12), it is observed that the results are in better agreement when the equivalent circular fin method is used, expected when the ratio XL/XT is 1.5, for which the underestimation more pronounced than with the sector method. For hexagonal fins (fig. 13), the equivalent circular fin method is in good agreement with the numerical model. For a ratio PL/PT of 2, the circular fin method overestimates the fin efficiency up to 3% while it is underestimated up to 4.5% with the sector method.

2.2 Louvered and slit fins

For enhanced fin designs with louvers and slits, as presented in figure 14, the fin shape alters the conduction path through the fin and the fin efficiency is expected to be overestimated by the 1-D sector method [9]. In this section, 2-D calculations are performed with constant heat transfer coefficient over the entire fin surface. It is observed in figure 15 presenting the temperature distribution within the fin, that the slits alter the temperature distribution. Figure 16 presents the fin efficiency calculations for the considered slit fin design. Firstly it is interesting to compare the slit fin efficiency to the plain fin efficiency, both numerically calculated. It is shown that the slit fin efficiency drops significantly compared to the plain fin efficiency due to the presence of slits which alter the conduction path and avoid fin temperature homogenization. In this case it is shown that the difference in fin efficiency is up to 4% in the practical range of m.XT parameter. Concerning the sector method applied to slit (or louver) fin efficiency calculation, in the case presented in figure 14, the calculation are performed for a small PL/PT ratio (0.63), so that the sector method is in very good agreement with the plain fin efficiency numerically calculated. Consequently, for this PL/PT ratio, the sector method clearly overestimates the slit fin efficiency (up to 5% in the considered range of m.XT parameter).

Figure 14: slit fin

Figure 15: Temperature distribution (K) within slit or louvered fin m.XT = 0.728 ; PL/PT = 0.63 ; PT/2r = 2.5 ; h = 40W/m2.K ; f = 170W/m.K ; f = 0.1mm Ta = 27C ; Tbase = 54C ; no. slit per slit zone NS = 5 ; slit depth SS / PL = 0.08

International Congress of Refrigeration 2003, Washington, D.C. 6

1,01 1 Fin efficiency ratios / 0,99 0,98 0,97 0,96 0,95 0,94 0 0,5 1 m XT 1,5 2 Fin efficiency ratios /

1,01 1 0,99 0,98 0,97 0,96 0,95 0,94 0 0,5 1 m XT 1,5 2 Numeric Slit / Numeric Plain Numeric Slit / Circular method Numeric Plain / Circular method

Numeric Slit / Numeric Plain Numeric Slit / Sector method Numeric Plain / Sector Method

Figure 16: Ratio of hexagonal fin efficiencies with and without slits calculated by numerical 2D model and sector method

Figure 17: Ratio of hexagonal fin efficiencies with and without slits calculated by numerical 2D model and equivalent circular fin method (Schmidt)

It is shown in figure 11 that for higher PL/PT ratios (>1.5), the sector method underestimates the plain fin efficiency. The magnitude of the difference in this case is comparable to the magnitude of the difference due to the presence of louvers and slits. Therefore, it could be expected that for louvered or slit fins with a high PL/PT ratio, the error generated by the sector method compensates the presence of slits, leading to a good agreement between the sector method and the slit fin efficiency. Concerning the equivalent circular fin method, for this small PL/PT ratio, the equivalent circular fin method (fig. 17) gives the same results as the sector method. The slit efficiency is overestimated by the circular fin method up to 5% in the considered range of m.XT parameter. Nevertheless, it is shown in figure 13, than the equivalent circular fin method overestimates the plain fin efficiency when the PL/PT ratio is increased (>1.5), and consequently when it is applied to slit or louvered fins, no good compensation could occur. Therefore the equivalent circular fin method always overestimates the slit / louver fin efficiency.

CONCLUSION
A review and a comparison of the usual analytical methods used for fin efficiency calculation in continuous fin-andtube heat exchangers have been performed. 2-D numerical models of rectangular and hexagonal fins have been performed in order to quantify the 1-D assumption validity used in the sector method and equivalent circular fin method. The sector method is in good agreement with the fin efficiency numerically calculated when the longitudinal tube pitch to transversal tube pitch ratio is near 1. When this ratio is higher, the sector method underestimates the fin efficiency. The same trends are shown for the equivalent circular fin method, but the results are in better agreement and it is preferable to use the equivalent circular fin method in the case of plain fins. In the case of enhanced fins having louvers or slits, it is shown that the fin efficiency is lower. Consequently, when the longitudinal tube pitch to transversal tube pitch ratio is near 1, the equivalent circular fin and sector methods overestimate significantly the enhanced fin efficiency. Nevertheless, when the longitudinal tube pitch to transversal tube pitch ratio is higher, the error introduced by the sector method is comparable to the effect of the presence of slits. Consequently the sector method is in good agreement with the slit fin computation in the case of high PL/PT ratios. This error compensation effect does not occur with the equivalent circular fin method.

International Congress of Refrigeration 2003, Washington, D.C. 7

NOMENCLATURE
Roman letter symbols Ai h In Kn m PT PL r rf rf,eq Ta Tbase XT XL XD area of sector i (m2) heat transfer coefficient (W/m2.K) Modified Bessel function of 1st kind Modified Bessel function of 2nd kind fin parameter defined by eq. (2) transversal tube pitch (m) longitudinal tube pitch (m) tube radius (m) circular fin radius (m) equivalent circular fin radius (m) surrounding air temperature (C ; K) fin base temperature (C ; K) half transversal tube pitch (m) (see figs. 3,4) half longitudinal tube pitch (m) (see figs. 3,4) distance defined by eq. (6) (m) Greek letter symbols f m f f i sector circular numeric fin thickness (m) parameter defined by eq. (3) modified parameter defined by eq. (4) thermal conductivity of fin (W/m.K) fin efficiency fin efficiency fin efficiency of sector i fin efficiency calculated by sector method fin efficiency calculated by equivalent circular fin method (Schmidt) fin efficiency calculated by 2-D num. model

REFERENCES
Wang, C.C., Lee, W.S., Sheu, W.J., 2001, A comparative study of compact enhanced fin-and-tube heat exchangers, International Journal of Heat and Mass transfer, vol. 44: p.3565-3573. 2. Du, Y.J., Wang, C.C., 2000, An experimental study of the air-side performance of the superslit fin-and-tube heat exchangers, International Journal of Heat and Mass Transfer, vol. 43: p.4475-4482. 3. Kim, N.H., Youn, B., Webb, R.L., 1999, Air side heat transfer and friction correlations for plain fin-and-tube heat exchangers with staggered tube arrangements, ASME Transaction, vol.121, august 1999. 4. Wang, C.C., Webb, R.L., Chi, K.Y., 2000, Data reduction for air-side performance of fin-and-tube heat exchangers, Experimental Thermal and Fluid Science, vol. 21: p.218-226. 5. Gardner, K.A., 1945, Efficiency of extended surface, Trans. ASME, Journal of Heat Transfer, no. 65: p.621631. 6. Incropera F.P., DeWitt D.P., 1996, Fundamental of heat and mass transfer 4th Edition, John Wiley & Sons. 7. Kraus, A.D., Aziz, A., Welty, J., 2001, Extended Surface Heat Transfer, John Wiley & Sons. 8. Schmidt, T.E., 1945, La production calorifique des surfaces munies dailettes, Bulletin de lInstitut International du Froid, Annexe G-5. 9. Hong, K.T., Webb, R.L., 1996, Calculation of fin efficiency for wet and dry fins, HVAC&R Research, vol. 2, no. 1: p.27-41. 10. Schmidt, T.E., 1949, Heat transfer calculation for extended surfaces, Refrigerating Engineering, vol. 57: p.351357. 1.

CALCUL DEFFICACITE DAILETTE AMELIOREE DANS LES ECHANGEURS DE CHALEUR EN REGIME SEC
RESUME : Le calcul de lefficacit des ailettes pour les changeurs tubes et ailettes air / fluide est de la plus grande importance pour lvaluation de la performance thermique de lchangeur ou pour la dtermination du coefficient dchange ct air partir de donnes exprimentales. Les changeurs compacts haute efficacit nergtique sont composs dailettes amliores ( persiennes ou fentes) pour lesquelles le calcul de lefficacit dailette peut tre surestim par les mthodes analytiques mono-dimensionelles telles que la mthode dite sectorielle. Lhypothse de flux conductif uniquement radial au sein de lailette est mise en dfaut par les persiennes ou les fentes qui perturbent la conduction dans lailette et lefficacit ne peut tre dtermine quen rsolvant numriquement lquation de la chaleur en 2 dimensions. Cela est effectu avec le logiciel SimTherm dvelopp par le Centre dEnergtique et les rsultats sont compars aux formulations classiques de calcul defficacit dailette.

International Congress of Refrigeration 2003, Washington, D.C. 8