Lei Zhang

Defu Che1

e-mail: dfche@mail.xjtu.edu.cn

State Key Laboratory of Multiphase

Flow in Power Engineering,

School of Energy and Power Engineering,

Xi’an Jiaotong University,

Xi’an 710049, China

An Experimental and NumericalInvestigation on theThermal-Hydraulic Performanceof Double Notched PlateThe double notched (DN) plate is commonly used in rotary air preheaters, but relevantinvestigations are rare. Thus, thermal-hydraulic performances of the DN plate are inves-tigated in this paper. A single-blow, transient technique is refined and then used to mea-sure the overall mean heat transfer coefficients and friction factors. A validatednumerical method is also utilized to provide local information. The measured resultsshow that the performance of the DN plate approaches that of the double undulated (DU)plate and lies between that of the cross corrugated (CC) plate and the parallel plate. Noswirling flow pattern is identified in the predicted velocity fields. Basically, two types offlow are observed: wavy channel flow and pipe flow. High or low Nusselt numbers, Nu,are obtained at the luff or lee side of undulations and notches, respectively. Nu valuesincrease and Nu distributions become more homogenous with increasing Reynolds num-bers, Re. A recommendation is made that the DN plate be operated under moderate Re toachieve homogenous and enhanced heat transfer, given the allowable pressure drop.[DOI: 10.1115/1.4006210]

Keywords: rotary air preheater, double notched plate, single-blow technique, numericalsimulation, heat transfer, friction factor

1 Introduction

A rotary air preheater is an important component of a fossil fuelpower plant because it performs the essential function of recover-ing low grade heat from the combustion product gases [1]. It typi-cally consists of many closely packed heat transfer elements. Itsmain performance requirements are high heat transfer rate, lowpressure loss, and rapid and accurate detection of hot spots andearly fires. These depend crucially on the geometrical design oftheir elements [2]. Some popular shapes are DN plate, CC plate,corrugated undulated plate, and DU plate [3]. Among them, theDN and CC plates are the most commonly used plate types.

The present DN plate is comprised of notches, undulations, andflat sections, see Fig. 1(b). The notches are longitudinally straight,equidistantly laterally spaced apart, and parallel to each other.Alternating between the notches is mutually parallel flat sections.Between the notches and flat sections are skewed undulations.The top, middle, and bottom DN plates in Fig. 1(a) are identicalto each other. This is advantageous in that only one type of plateneeds to be manufactured. Undulations of the DN plate introduceboundary layer interruptions and therefore improve heat transfer.The channels between DN plates are generally longitudinallystraight, providing a line of sight view through them for efficientdetection of hot spots and element fires.

The CC plate is also widely used for its relatively high thermalperformance, low pressure drop, simple structure, and high rigid-ity [4]. A large number of studies [5–13] have been carried outto investigate the performance of the CC plate, while only alimited amount of data exists for the DN plate because of com-mercial restrictions. To the best of our knowledge, systematic

investigation of thermal-hydraulic performance of the DN platehas not been reported in the literature.

Therefore, a single-blow technique (SBT) is conducted to testthe overall performance of the DN plate in this paper. In addition,a numerical simulation is carried out to provide the detaileddescription of the flow and temperature field, which is very diffi-cult to be achieved experimentally and is helpful for better under-standing of the flow and heat transfer mechanisms. In Fig. 1(c),four typical sections, L–L, M–M, N–N, and O–O, are illustratedas slices of the computational domain. Numerical results havebeen presented for these sections.

2 Experimental Method

2.1 Introduction of Single-Blow Technique. Experimentaltechniques for obtaining heat transfer performance of heatingsurfaces can be classified into the steady-state method [14,15] andthe transient method. The transient method is composed of threeelements: an experiment in which a heat transfer matrix is oper-ated as a regenerator, a model that links measured fluid tempera-tures to the heat transfer coefficient, and an evaluation scheme bywhich the measured data can be compared with the prediction ofthe model [16]. The SBT is a relatively simple transient method,which employs a single fluid. In SBT, the inlet fluid temperatureis varied as a function of time, and the resulting exit fluid tempera-ture history is measured. The measured outlet fluid temperatureis then matched with the solutions of an analytical model todetermine the average heat transfer coefficient [17]. The SBT hasseveral advantages over the steady-state method: The apparatus isrelatively simple and the experimental time for each data point isrelatively short. More importantly, the rather difficult task ofmeasuring the complicated DN plate temperature in the steady-state method is avoided in the SBT. Consequently, the SBT is pre-ferred in this paper to test the DN plate.

1Corresponding author.Contributed by the Heat Transfer Division of ASME for publication in the

JOURNAL OF HEAT TRANSFER. Manuscript received June 14, 2011; final manuscriptreceived February 13, 2012; published online July 2, 2012. Assoc. Editor: Wei Tong.

Journal of Heat Transfer SEPTEMBER 2012, Vol. 134 / 091802-1Copyright VC 2012 by ASME

Downloaded 05 Jan 2013 to 117.32.153.173. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

The original mathematical model and its analytical solutions ofthe single-blow problem were developed by Schumann [18]. Hissolutions were first used as the basis for the transient technique byFurnas [19]. Pucci et al. [20] presented an excellent summary ofthe underlying theory of the SBT. Liang and Yang [17] developeda modified SBT to minimize the errors caused by assumption ofstep change inlet fluid temperature. Based on numerical simula-tion, Mullisen and Loehrke [21] extended Liang–Yang analysis.Their method has been accepted as one of the best methods forreducing transient test data [22]. The SBT developed by Mullisenand Loehrke is, therefore, utilized in this paper.

2.2 Mathematical Model. The accuracy of SBT mainlydepends upon how precisely the model describes the experimentalprocess. Cheng and Huang [23] found that neglecting the effectsof longitudinal conduction and thermal capacity of fluid maycause considerable error. Thus, Mullisen–Loehrke method is fur-ther refined in this paper to include these effects.

Assumptions made in the current analysis are as follows:

(1) Thermophysical properties of the fluid and test core are in-dependent of temperature.

(2) The fluid has steady flow and is uniformly distributed overthe matrix cross section.

(3) Thermal conductivity of both fluid and solid is infinite per-pendicular to the flow direction.

(4) Thermal conductivity of both fluid and solid is finite in theflow direction.

(5) The boundaries of the solid test core are adiabatic.

The test core is shown in Fig. 2. The method provided bySheer et al. [24] is used to define the control volume. The energyconservation equations for an element dx can be summarized asfollows:

qsdscs

@Ts1

@t¼ ksds

@2Ts1

@x2þ h Tf � Ts1ð Þ (1)

qsdscs

@Ts2

@t¼ ksds

@2Ts2

@x2þ h Tf � Ts2ð Þ (2)

qfdfcf

@Tf

@tþ qfufdfcf

@Tf

@t¼ kfdf

@2Tf

@x2þ h Ts1 þ Ts2 � 2Tfð Þ (3)

The above governing equations are subject to the following initialand boundary conditions:

Ts1 x; t ¼ 0ð Þ ¼ Ts2 x; t ¼ 0ð Þ ¼ Tf x; t ¼ 0ð Þ ¼ T0 (4)

Tf x ¼ 0; tð Þ ¼ Tin tð Þ (5)

Fig. 2 (a) DN plate test core and (b) energy balance inside acontrol volume

Fig. 1 The DN plate geometry: (a) end view, (b) perspective view, and (c) top view

091802-2 / Vol. 134, SEPTEMBER 2012 Transactions of the ASME

Downloaded 05 Jan 2013 to 117.32.153.173. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

Tf x ¼ L; tð Þ ¼ Tout tð Þ (6)

@Ts1 x ¼ 0; tð Þ@x

¼ @Ts1 x ¼ L; tð Þ@x

¼ @Ts2 x ¼ 0; tð Þ@x

¼ @Ts2 x ¼ L; tð Þ@x

¼ 0 (7)

where Tin(t) and Tout(t) are the measured inlet and outlet fluid tem-peratures, respectively.

These equations are converted to finite-difference equations bycontrol volume integration method [25]. The resultant discreteequations are as follows:

qsdscsDx2 þ 2ksdsDtþ hDtDx2

qsdscsDx2

� �Tnþ1

s1;i

¼ Tns1;i þ

ksdsDt

qsdscsDx2Tnþ1

s1;iþ1 þ Tnþ1s1;i�1

� �þ hDtDx2

qsdscsDx2

� �Tnþ1

f;i

(8)

qsdscsDx2 þ 2ksdsDtþ hDtDx2

qsdscsDx2

� �Tnþ1

s2;i

¼ Tns2;i þ

ksdsDt

qsdscsDx2Tnþ1

s2;iþ1 þ Tnþ1s2;i�1

� �þ hDtDx2

qsdscsDx2

� �Tnþ1

f;i

(9)

qfdfcfDx2 þ qfufdfcfDtDxþ 2kfdfDtþ 2hDtDx2� �

Tnþ1f;i

¼ qfufdfcfDtDxþ kfdfDtð ÞTnþ1f;i�1 þ hDtDx2

� �Tnþ1

s1;i þ Tnþ1s2;i

� �

þ kfdfDtð ÞTnþ1f;iþ1 þ qfdfcfDx2

� �Tn

f;i (10)

The discrete initial and boundary conditions are

T1s1;i ¼ T1

s2;i ¼ T1f;i ¼ T0 (11)

Tnf;0 ¼ Tin tð Þ (12)

Tnf;mþ1 ¼ Tout tð Þ (13)

Tns1;0 � Tn

s1;1 ¼ Tns1;m � Tn

s1;mþ1 ¼ Tns2;0 � Tn

s2;1 ¼ Tns2;m � Tn

s2;mþ1 ¼ 0

(14)

The discrete equations are solved using the tridiagonal matrixalgorithm described in Ref. [26].

A unique procedure is developed to obtain the average convec-tion coefficient, h: guess at h first, solve Eqs. (8)–(10) subject toconditions (11)–(14). The predicted temperature history is thencompared with Tout(t). The value of h is iteratively changed untilthe two temperature histories match within specific limits.

2.3 Experimental Apparatus and Procedures. Figure 3shows a schematic diagram of the experimental apparatus. It is

basically an open circuit subsonic wind tunnel with a300� 300� 900 mm3 test section. Honeycomb straighteners andfine wire gauze were installed to ensure steady and uniform airflow into the test section. An electric resistance heater is installedto provide a 40 K rise in incoming air temperature. By limitingthe temperature difference to 40 K, the thermophysical propertiesof the fluid and the test core are essentially constant. Elastomericinsulation is put around the wind tunnel to achieve the adiabaticboundary condition.

Two thermocouple grids are placed at the inlet and outlet of thetest section to record air temperature. The uncertainty in tempera-ture measurement is within 60.5 K. Static pressure taps arelocated upstream and downstream of the test core. Two standardorifices with different ranges were used to improve measuringaccuracy of the mass flow rate. All the differential pressure signalsare converted to electric signals by pressure transducers with anaccuracy of 0.5%.

The data acquisition system is activated at first to record tem-perature history. Electrical supply to the heaters is switched ononce temperatures of the fluid and the DN plates reach steadystate. The air temperature is continuously recorded at a samplingrate of 1 Hz for a period of about 600 s. The pressure drops acrossthe orifice plate and the test core are also measured. By conduct-ing numerous tests at different flow rates, j and f data could bedetermined for various Re.

2.4 Data Reduction and Discussion. The heat transferresults are presented in terms of the Nusselt number or Colburnfactor, j. The Colburn factor is defined as

j ¼ Nu

RePr0:5¼ hD

kf

1

RePr0:5f

(15)

Friction loss is presented in terms of the Darcy friction factor f

f ¼ D

L

2Dp

qfu2f

� ðKc þ KeÞ�

(16)

where Kc and Ke are pressure loss coefficients at the inlet and out-let of the test core, respectively [14].

The root-sum-square method described by Moffat [27] is usedto analyze the uncertainties of the experimental results. The uncer-tainties of j and f are estimated to be about 4.6% and 5.5%,respectively.

Figure 4 shows typically measured and calculated outlet tem-peratures. In the figure, the theoretical exit air temperature bestmatches the experimental curve at h¼ 73.87 W/m2K, and, the cal-culated root-mean-square deviation between the theoretical andexperimental curve is 0.42 �C. Measured j and f values are pre-sented in Figs. 5 and 6, together with the results of some typicalplates in literature. These plates are the CC plate in Ref. [5], theDU plate in Ref. [24], and the parallel plate in Ref. [28]. The datafor CC plate are scaled to the right axis, while other data are

Fig. 3 Test apparatus

Journal of Heat Transfer SEPTEMBER 2012, Vol. 134 / 091802-3

Downloaded 05 Jan 2013 to 117.32.153.173. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

scaled to the left axis. With pure skew undulations on it, the CCplate introduces efficient interruptions of the boundary layers,and, thus, gains good thermal performance with a large pressuredrop penalty. On the contrary, the parallel plate gains poor ther-mal performance with a small pressure drop. Performances of thesimilar DN and DU plates lie well between that of the CC plateand the parallel plate. A conclusion, therefore, can be drawn that acompromise has been made between high heat transfer rate, effi-cient fault detection, and effective soot blowing by adding notchesand flat sections to the angled undulations.

3 Numerical Simulations

3.1 Computational Grid and Model. The unstructured Tet/Hybrid grid presented in Fig. 7 is generated for the current DNplate. According to the grid independency check, the total of1,396,578 grid cells used in this paper is sufficient.

In a comparison study, Patel et al. [29] found that the low-Reynolds number k� e (LBKE) model developed by Lam andBremhorst [30] gives good performance for channel flows. Subse-quently, Ciofalo et al. [10] confirmed that the best performance isfrom the LBKE model, among five numerical models for model-ing CC channel flows. Since no suggested model is available inthe literature for the DN plate, the LBKE model is used in thepresent study.

3.2 Solution Methodology. Uniform inlet velocity profilesand uniform inlet temperature distributions are assumed, whilezero-normal derivative conditions are imposed on the outlet faces.The calculation proceeds along the flow direction until fully

Fig. 5 Current versus previous j data

Fig. 6 Current versus previous f data

Fig. 7 Unstructured grids for a DN plate

Fig. 8 Experimental and numerical j and f values

Fig. 4 Experimental data processing for Re 5 4709

Fig. 9 In-plane velocity fields in the section N–N

091802-4 / Vol. 134, SEPTEMBER 2012 Transactions of the ASME

Downloaded 05 Jan 2013 to 117.32.153.173. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

developed situations are achieved. This is done by repeatedly set-ting outlet distributions as inlet conditions (see more details inRef. [10]). No-slip conditions and uniform wall temperature con-ditions are used. Air is the working fluid and its density, specificheat, thermal conductivity, and viscosity are taken as second-order polynomial functions of its temperature.

The Navier–Stokes equations are solved by the SIMPLECscheme, the convective terms are solved by the third-orderQUICK scheme, and the diffusive terms by a second-order centraldifference scheme. The residuals, together with the outlet pres-sure, velocity, and temperature, are monitored to identifyconvergence.

3.3 Results and Discussion. Numerical versus experimentalj and f values are shown in Fig. 8. The predicted deviations forj and f are from �9.25% to 11.38% and from �17.46% to13.99%, respectively. Such a level of agreement is regarded assatisfactory and confirms the validity of the numerical method. Asimilar statement was reported by Song and Nishino [31] whoconducted single-blow tests and simulations on a fin-channel andclaimed that the measured heat transfer coefficients matched wellwith those of the steady-state computation.

Similar velocity fields are identified for different Re in Fig. 9.The swirling flow pattern, as was identified in CC channel byCiofalo et al. [10], is invisible in the present DN channel. Theinteraction between the two sets of fluid streams flowing along theupper and lower undulations is greatly reduced in the DN channelas the undulations are separated apart at a certain distance by thedeep notches. Thus, no swirling flow occurs and, consequently,both the pressure drops and heat transfer rates in DN channel aresmaller than those in the CC channel.

The velocity distribution feature is evident in Fig. 10: Intensesecondary flows appear in the undulated region, while secondaryflows are hardly visible in the notched region. Flow and heattransfer processes in the notched region are similar to those of thepipe flow. In the undulated region, fluid streams near the troughsof the top undulations are strongly affected by the crossingstreams along the middle undulations. It is evident in Fig. 10(a)that part of the fluid streams near the troughs of the top undula-tions joins the streams along the middle undulations. Such stronginteraction is believed to account for fluid shift and enhanced heattransfer. With increasing Re, the thickness of thermal boundarylayer decreases, as shown in Fig. 10(b), which leads to heat trans-fer enhancement. Generally speaking, a conclusion can be drawnfrom Figs. 9 and 10 that Re does not significantly influence theflow patterns and temperature distributions in the DN channel.

More prediction results are provided in Fig. 11. Fluid flow canbe characterized as wavy channel flow and pipe flow for sectionsL–L and M–M, respectively. In section L–L, the boundary layer iseffectively disturbed. As regards section M–M, pipe flow featuresare easily identified. It is obvious in Fig. 11(b) that the thermal

Fig. 10 (a) In-plane velocity fields and (b) temperature fields inthe section O–O

Fig. 11 (a) In-plane velocity fields and (b) the correspondingtemperature fields

Fig. 12 Distributions of the Nusselt number on the bottom wall

Journal of Heat Transfer SEPTEMBER 2012, Vol. 134 / 091802-5

Downloaded 05 Jan 2013 to 117.32.153.173. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

boundary layer in section L–L is thinner than that in M–M, indi-cating better heat transfer performance in the wavy channel.

Figure 12 reports predicted Nu distributions on the bottom wallof the computational domain. High Nu (light color) is observed atthe luff side of the undulations and the notches where the flowimpinges. Low Nu is obtained at the lee side of the undulationsand the notches and near the contact lines where notches meet flatsections. Similar experimental data were reported for the CCchannel by Gaiser and Kottke [32]. Moreover, from Fig. 12, it canbe concluded that the Nu distribution becomes more homogenouswith increasing Re, implying strong mixing and enhanced heattransfer as Re increases.

Further detail is given by equally dividing the computationaldomain into 20 pieces in the x direction. A Nusselt number foreach piece is obtained and normalized by the maximum value,Numax. The normalized Nu distributions are presented in Fig. 13.At small Re values, a remarkable wave crest is identified near themiddle part of the DN plate. At such low Re, laminar flow isbelieved to prevail in the whole channel. Therefore, sections withlarger flow rate will achieve higher Nu values. The pressure dropof the notched section is smaller than that of the undulated sectionat the same Re, for equal flow rates. Thus, more flow rate isdistributed to the notched channel than to the undulated channelto maintain the same pressure drop, which leads to the aforemen-tioned Nu curve crest. At large Re values, heat transfer enhance-ment effects in the undulated region prevail. As Re increases,normalized Nu values of the undulated section approach andfinally exceed the maximum Nu value of the notched section. Itcan be concluded that heat transfer is intensified and the intensiveheat transfer center shifts from the notched sections to the undu-lated sections with increasing Re. Thus, operation under moderateRe values is recommended for the present DN plate to achieve ho-mogenous heat transfer given the allowable pressure drop.

4 Conclusions

A refined, single-blow transient experimental technique, alongwith a validated numerical method, is used to investigate the com-plex flow and heat transfer phenomena within the DN channel.

Measured j and f are presented and compared with the literaturedata over a wide Re range. It is found that the thermal-hydraulicperformance of the DN plate lies between that of the CC plate andthe parallel plate, which confirms the statement that heat transferand energy recovery have been compromised in the DN plate toprovide effective soot blowing and fault detection. Moreover,comparable j and f are observed for the similar DN and DU plates.

The current numerical method is validated by good agreementbetween the experimental and numerical data. Predicted velocityfields show no swirling flow patterns. Basically, two distinct flowtypes are identified in the DN channel, i.e., the wavy channel flowand the pipe flow. The former is believed to be the main reasonfor heat transfer enhancement under medium and high Re flow.

High Nu is obtained at the luff side of the undulations and thenotches, while low Nu is obtained at the lee side of the undula-tions and the notches. In addition, it is found that the Nu distribu-tion becomes more homogenous and Nu values increase withincreasing Re. Thus, it is recommended that the DN plate shouldbe operated under moderate Re values to achieve homogenousand enhanced heat transfer while maintaining allowable pressuredrop.

These are useful information and may serve as the starting pointfor further research about optimal design of DN plates.

NomenclatureA ¼ areac ¼ specific heat

D ¼ hydraulic diameterf ¼ Darcy friction factorh ¼ heat transfer coefficientj ¼ Colburn factork ¼ turbulence energy or thermal conductivity

Kc, Ke ¼ entrance and exit pressure loss coefficientsL ¼ total lengthm ¼ mass_m ¼ mass flow rate

Nu ¼ Nusselt numberPr ¼ Prandtl numberDp ¼ pressure dropRe ¼ Reynolds number

t ¼ timeDt ¼ time intervalT ¼ temperatureu ¼ reference velocity

x,y,z ¼ Cartesian coordinatesDx ¼ length intervald ¼ thicknesse ¼ dissipation rate of turbulence energyh ¼ included angleq ¼ density

Subscriptscond ¼ conductionconv ¼ convection

f ¼ fluidi ¼ section number

in ¼ inletmax ¼ maximumout ¼ outlet

s ¼ solids1 ¼ the first DN plates2 ¼ the second DN plate

Superscriptn ¼ time layer number

References[1] Chew, P. E., 1985, “Rotary Air Preheaters on Power Station Boilers,” Proceed-

ings of the Symposium on Waste Heat Recovery and Utilisation, Institute ofEnergy, UK.

[2] Stasiek, J., Collins, M. W., Ciofalo, M., and Chew, P. E., 1996, “Investigationof Flow and Heat Transfer in Corrugated Passages—I. Experimental Results,”Int. J. Heat Mass Transfer, 39(1), pp. 149–164.

[3] Ciofalo, M., Collins, M. W., and Stasiek, J. A., 1998, Flow and Heat TransferPredictions in Flow Passages of Air Preheaters: Assessment of AlternativeModeling Approaches, Computational Mechanics Publications, Southampton,UK, pp. 169–225.

[4] Zhang, L., and Che, D., 2011, “Influence of Corrugation Profile on the Thermal-hydraulic Performance of Cross-Corrugated Plates,” Numer. Heat Transfer,Part A, 59(4), pp. 267–296.

[5] Focke, W. W., Zachariades, J., and Olivier, I., 1985, “The Effect of the Corru-gation Inclination Angle on the Thermohydraulic Performance of Plate HeatExchangers,” Int. J. Heat Mass Transfer, 28(8), pp. 1469–1479.

Fig. 13 Spanwise Nusselt number distributions

091802-6 / Vol. 134, SEPTEMBER 2012 Transactions of the ASME

Downloaded 05 Jan 2013 to 117.32.153.173. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

[6] Muley, A., and Manglik, R. M., 1999, “Experimental Study of Turbulent FlowHeat Transfer and Pressure Drop in a Plate Heat Exchanger With ChevronPlates,” ASME J. Heat Transfer, 121(1), pp. 110–117.

[7] Wang, Q. W., Zhang, D. J., and Xie, G. N., 2009, “Experimental Study andGenetic-Algorithm-Based Correlation on Pressure Drop and Heat Transfer Per-formances of a Cross-Corrugated Primary Surface Heat Exchanger,” ASME J.Heat Transfer, 131(6), p. 061802.

[8] Muley, A., Manglik, R. M., and Metwally, H. M., 1999, “Enhanced Heat Trans-fer Characteristics of Viscous Liquid Flows in a Chevron Plate HeatExchanger,” ASME J. Heat Transfer, 121(4), pp. 1011–1017.

[9] Rao, B. P., Sunden, B., and Das, S. K., 2005, “An Experimental and TheoreticalInvestigation of the Effect of Flow Maldistribution on the Thermal Performanceof Plate Heat Exchangers,” ASME J. Heat Transfer, 127(3), pp. 332–343.

[10] Ciofalo, M., Stasiek, J., and Collins, M. W., 1996, “Investigation of Flow andHeat Transfer in Corrugated Passages—II. Numerical Simulations,” Int. J. HeatMass Transfer, 39(1), pp. 165–192.

[11] Jain, S., Joshi, A., and Bansal, P. K., 2007, “A New Approach to NumericalSimulation of Small Sized Plate Heat Exchangers With Chevron Plates,”ASME J. Heat Transfer, 129(3), pp. 291–297.

[12] Zhang, L. Z., 2005, “Turbulent Three-Dimensional Air Flow and Heat Transferin a Cross-Corrugated Triangular Duct,” ASME J. Heat Transfer, 127(10), pp.1151–1158.

[13] Blomerius, H., Holsken, C., and Mitra, N. K., 1999, “Numerical Investigationof Flow Field and Heat Transfer in Cross-Corrugated Ducts,” ASME J. HeatTransfer, 121(2), pp. 314–321.

[14] Kays, W. M., and London, A. L., 1984, Compact Heat Exchangers, McGraw-Hill, New York.

[15] Kays, W. M., and London, A. L., 1950, “Heat Transfer and Flow Friction Char-acteristics of Some Compact Heat Exchanger Surfaces,” Trans. ASME, 72, pp.1075–1097.

[16] Loehrke, R. I., 1990, “Evaluating the Results of the Single-Blow Transient HeatExchanger Test,” Exp. Therm. Fluid Sci., 3(6), pp. 574–580.

[17] Liang, C. Y., and Yang, W. J., 1975, “Modified Single-Blow Technique for Per-formance Evaluation on Heat Transfer Surfaces,” ASME J. Heat Transfer,97(1), pp. 16–21.

[18] Schumann, T. E. W., 1929, “Heat Transfer: A Liquid Flowing Through a Po-rous Prism,” J. Franklin Inst., 28(1), pp. 405–416.

[19] Furnas, C. C., 1932, “Heat Transfer From a Gas Stream to a Bed of Broken Sol-ids,” US Bureau of Mines Bulletin No. 361.

[20] Pucci, P. F., Howard, C. P., and Piersall, C. H., Jr., 1967, “The Single-BlowTransient Testing Technique for Compact Heat Exchanger Surfaces,” ASME J.Eng. Power, Series A, 89(2), pp. 29–40.

[21] Mullisen, R. S., and Loehrke, R. I., 1986, “A Transient Heat Exchanger Evalua-tion Test for Arbitrary Fluid Inlet Temperature Variation and LongitudinalCore Conduction,” ASME J. Heat Transfer, 108(2), pp. 370–376.

[22] Krishnakumar, K., John, A. K., and Venkatarathnam, G., 2011, “A Reviewon Transient Test Techniques for Obtaining Heat Transfer Design Data ofCompact Heat Exchanger Surfaces,” Exp. Therm. Fluid Sci., 35(4), pp.738–743.

[23] Cheng, C., and Huang, C., 1994, “Extended Model for Single-Blow TransientTesting Method in Evaluating Thermal Performance of Heat TransferSurfaces,” Int. Commun. Heat Mass Transfer, 21(1), pp. 53–63.

[24] Sheer, T. J., De Klerk, G. B., Jawurek, H. H., and Lander, M., 2006, “A Versa-tile Computer Simulation Model for Rotary Regenerative Heat Exchangers,”Heat Transfer Eng., 27(5), pp. 68–79.

[25] Tao, W., 2002, Numerical Heat Transfer, Xi’an Jiaotong University Press,Xi’an, pp. 39–43.

[26] Tao, W., 1991, Computational Fluid Dynamics and Heat Transfer, ChinaArchitecture & Building Press, Beijing, pp. 25–27.

[27] Moffat, R. J., 1988, “Describing the Uncertainties in Experimental Results,”Exp. Therm. Fluid Sci., 1(1), pp. 3–17.

[28] Rohsenow, W. M., Hartnett, J. P., and Cho, Y. I., 1998, Handbook of HeatTransfer, McGraw-Hill, New York, Chap. V.

[29] Patel, V. C., Rodi, W., and Scheuerer, G., 1985, “Turbulence Models for Near-Wall and Low Reynolds Number Flows: A Review,” AIAA J., 23(9), pp.1308–1319.

[30] Lam, C. K. G., and Bremhorst, K., 1981, “A Modified Form of the K-EModel for Predicting Wall Turbulence,” ASME J. Fluids Eng., 103(3), pp.456–460.

[31] Song, G.-D., and Nishino, K., 2008, “Conjugate Heat Transfer Computation forEvaluation of Single-Blow Method for Compact Fin-Tube Heat Exchangers,” J.Therm. Sci. Technol., 3(2), pp. 219–233.

[32] Gaiser, G., and Kottke, V., 1989, “Flow Phenomena and Local Heat and MassTransfer in Corrugated Passages,” Chem. Eng. Technol., 12(1), pp. 400–405.

Journal of Heat Transfer SEPTEMBER 2012, Vol. 134 / 091802-7

Downloaded 05 Jan 2013 to 117.32.153.173. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm