Thermal performance of tubular heat exchanger with multiple twisted-tape inserts☆

Suriya Chokphoemphun,Monsak Pimsarn,Chinaruk Thianpong,Pongjet Promvonge*

Department of Mechanical Engineering,Faculty of Engineering,King Mongkut's Institute of Technology Ladkrabang,Bangkok 10520,Thailand

Keywords:Thermal performance Twisted tape arrangement Turbulator Nusselt number Friction factor

A B S T R A C T The paper presents an experimental investigation on enhanced heat transfer and pressure loss characteristics by using single,double,triple,and quadruple twisted-tape inserts in around tube having a uniform heat-fluxed wall.The investigation has been conducted in the heat exchanger tube inserted with various twisted-tape numbers for co-and counter-twist arrangements for the turbulent air flow,Reynolds number(Re)from 5300 to 24000.The typical single twisted-tape inserts at two twist ratios,y/w=4 and 5,are used as the base case,while the other multiple twisted-tape inserts are at y/w=4 only.The experimental results of heat transfer and pressure drop in terms of Nusselt number(Nu)and friction factor(f),respectively,reveal that Nu increases with the increment of Re and of twisted-tape number.The values of Nu for the inserted tube are in a range of 1.15–2.12 times that for the plain tube while f is 1.9–4.1 times.The thermal enhancement factor of the inserted tube under similar pumping power is evaluated and found to be above unity except for the single and the double co-twisted tapes.The quadruple counter-twisted tape insert provides the maximum thermal performance.

1.Introduction

Heat exchanger as equipment to facilitate the convective heat transfer of fluid inside tubes is frequently utilized in many industrial applications,such as chemical engineering process,heat recovery,air conditioning and refrigeration systems,power plant and radiators for automobiles.In general,heat transfer enhancement in heat exchangers can be divided into two methods.One is the active method requiring extra external power sources such as fluid vibration,injection and suction of the fluid,jet impingement and electrostatic fields.The other is the passive method that requires no other power source.The devices in this category are surface coating,rough surfaces,turbulent/swirl flow devices,extended surfaces etc.

Several investigations have been carried out to study the effect of turbulators(turbulent promoters)with different geometries on thermal behaviors in the heat exchanger,for example twisted-tapes[1,2],wirecoils[3,4],dimpled or grooved tubes[5,6],winglet/ fins[7,8],and combined turbulators[9,10].However,twisted tapes as one of passive turbulators have been applied extensively to enhance convection heat transfer in heat exchanger systems due to the need for finding the way to reduce the size and cost of those systems.For decades,the heat transfer enhancement by twisted-tape insert has been widely investigated both experimentally and numerically.Eiamsa-ard et al.[1]reported the heat transfer and pressure loss behaviors in a double pipe heat exchanger fitted with regularly-spaced twisted-tape elements at several space ratios.Promvonge[9]conducted measurements using wire coil in conjunction with twisted tape for heat transfer augmentation and reported that this combination leads to a double increase in heat transfer over the use of wire coil/twisted tape alone.

Krishna et al.[11]experimentally investigated the heat transfer characteristics in a circular tube fitted with straight full twist insert with different spacer distances.Influence of the tube equipped with the short-length twisted tape on Nu,f and thermal performance characteristics for several tape-length ratios was examined by Eiamsa-ard et al.[12].The effect of twisted tape consisting wire-nails and plain twisted tapes with three different twist ratios fitted in a heat exchanger pipe using water as the test fluid on thermal characteristics was studied experimentally by Murugesan et al.[13].Liao and Xin[14]reported the heat transfer behaviors in a tube with three-dimensional internal extended surfaces and twisted-tape inserts with various working fl uids.Chiu and Jang[15]presented the experimental and numerical analyses on thermal–hydraulic characteristics of air flow inside a circular tube with 5 different tube inserts;longitudinal strip inserts both with/without holes and twisted-tape inserts with three different twist angles for inlet velocity ranging from 3 to 18 m·s?1.Eiamsa-ard and Promvonge[16]conducted an experimental study on turbulent flow and heat transfer characteristics in a tube equipped with two types of twisted tapes:(1)typical twisted tapes and(2)alternate clockwise and counterclockwise twisted-tapes.Nine different clockwise and counterclockwise twisted-tapes were tested in that work and included the tapes with three twist-ratios and three twist-angles.The experiments were performed for Reynolds number of 3000 to 27000 using water as working fluid.The twin and triple twisted tapes used to generate twin and triple swirl flows in a circular tube were reported by Chang et al.[17].

Hong et al.[18]performed a numerical simulation of turbulent flow and heat transfer in converging–diverging tubes and converging–diverging tubes equipped with twin counter-swirling twisted tapes.In their work,effect of Re,pitch length,rib height,gap distance between twin twisted-tapes and tape number on Nu,f and thermal enhancement factor(η)were investigated.A simulation of multilongitudinal vortices in a tube induced by multiple twisted-tapes inserts for the Re from 300 to 1800 was examined by Zhang et al.[19].Effect of the combined conical-ring turbulator and twisted tape on heat transfer characteristics and thermal performance was also studied by Promvonge and Eiamsa-ard[20].Bharadwaj et al.[21]investigated the heat transfer and pressure drop in a spirally grooved tube with twisted-tape insert for laminar to fully turbulent regions.The effect of the direction of twist(left-twist and right-twist)on the thermo-hydraulic characteristics was also reported.Date and Gaitonde[22]developed the correlations for predicting the heat transfer coefficient and friction loss of laminar flow in a tube fitted with regularly spaced twisted-tape elements.Rahimi et al.[23]examined the heat transfer and thermal hydraulic performance in a round tube fitted with classic and modified twisted-tape inserts.They found that the jagged twisted tape provides more heat transfer enhancement than other tapes.Eiamsa-ard et al.[24]numerically studied the convective heat transfer in a circular tube fitted with loose- fit twisted tapes.Liu et al.[25]conducted a numerical investigation on the effect of short-width twisted-tapes inserted in a circular tube on heat transfer behaviors in laminar and turbulent flow s.

In the literature review above,many investigations are almost focused on the use of single,double and triple twisted tapes with similar tape-twist direction,apart from the modified twisted-tape while the effect of the tape number,co-and counter-twist arrangements has rarely been reported.Therefore,the utilization of double,triple and quadruple twisted-tapes in various forms of co-and counter-twist arrangements is offered as an enhancement device in the present work.The experiments are performed using the left-and right-twisted tapes arranged in different forms of counter-twisted and co-twisted tape pairs with the Re range in turbulent region,where air is used as the test fluid.The effect of the twisted-tape number and the combined counter-and co-swirls on the heat transfer rate,f and thermal performance characteristics in a round tube is determined in the current work.

2.Experimental Setup

A detail of the experimental apparatus used is displayed schematically in Fig.1.In the apparatus setting above,inlet air at 25°C from a 1.5 kW blow er was directed through an orifice flow meter and passed in the heat transfer test section.The air flow rate was measured by the orifice meter,built according to ASME standard[26]and calibrated by using a hot-wire anemometer to measure flow velocities across the tube section.Manometric fluid was used in an inclined-manometer with specific gravity(SG)of 0.826 to ensure reasonably accurate measurement of pressure drop across the orifice.The volumetric air flow rate from the blow er was adjusted by varying the motor speed of the blower through an inverter.The inner(D)and outer diameters of the copper test tube were,respectively,50.8 and 54.8 mm,and the tube was 3000 mm long,including the test section of L=1000 mm located at the exit.The test tube was heated by continually winding flexible electrical wire on the outer tube wall.The electrical output power was controlled by a Variac transformer to obtain a uniform heat- flux along the entire length of the test section.The outer surface of test tube was well insulated to minimize heat loss to the surrounding.The inlet and outlet air temperatures in the test tube were measured by resistance temperature detector thermocouples(RTD-Pt100)with±0.1°C accuracy which are positioned around 50 mm upstream and downstream of the test tube while the surface temperatures(Tw)were measured by 12 K-type thermocouples located equally on both top and side walls along the test section.All 24 K-type thermocouples were embedded in holes drilled from the outer surface and centered in the tube walls with the respective junctions positioned within 1 mm of the inside wall,and axial separation of the thermocouples was 90 mm apart and calibrated within±0.2°C deviation by thermostat before being used.All of the temperatures getting from the system were consistently recorded using a data logger.The pressure drop across the test section for calculation of friction factor was measured at an isothermal condition by a digital manometer with an accuracy of±0.5%.Reynolds numbers for air flowing through the test section were controlled in the range of 5300 to 24000.

Fig.2 shows the photographs of the test tube inserted with various numbers and arrangements of twisted tapes.In the figure,details on twisted-tape insertions indicating the number of twisted tapes(N)and the arrangement and the designation of each case were provided.All twisted-tapes were inserted in the test section with slightly tight fit and multiple twisted-tapes were attached together by superglue before insertion.The twisted-tapes made of aluminum sheet were 1200 mm long and 0.8 mm thick.The typical single twisted-tape had a width(w)of 42 mm with two different twist lengths(y):168 and 210 mm(twist ratio,y/w=4 and 5)while the rest were 21 mm wide with 84 mm twist-length(y/w=4).All tapes were twisted in two different directions:left-twist(LT)and right-twist(RT),and were arranged in 8 different configurations(2T1,2T2,3T1,3T2,4T1,4T2,4T3 and 4T4)as can be seen in Fig.2.

Fig.1.Schematic diagram of experimental apparatus.

Fig.2.Test tube inserted with various numbers of twisted tapes.

The uncertainty calculation based on Ref.[27]reveals that the maximum uncertainties of non-dimensional parameters were±5%for Reynolds number,±7.6%for Nusselt number and±9.5%for friction factor.The uncertainty in the axial velocity measurement was estimated to be less than±5%,and pressure has a corresponding estimated uncertainty of±5%,where as the uncertainty in temperature measurement at the tube wall was about±0.5%(±0.35 °C).

3.Data Reduction

The aim of the current work is to determine the heat transfer rate in a circular tube fitted with multiple twisted-tapes.The independent parameters are Reynolds number(Re),the number of twisted tapes(N)and twisted-tape arrangements.The Reynolds number is given by

The friction factor(f)computed by pressure drop across the test tube length(L)is written as

in which U is the mean air velocity in the test tube.

In the experiment,air fl owed through the test tube under a uniform heat- flux condition.The steady state of the heat transfer rate means that the heat transferred across the test section wall Qconvis equal to the heat absorbed by flowing air Qa:

with

and the convection heat transfer from test section can be written by

w here Twis local wall temperature and evaluated at the outer wall surface of test tube.The average wall temperature was calculated from 24 points of surface temperatures lined equally between the inlet and the exit of test tube.The average heat transfer coefficient(h)and mean Nusselt number(Nu)are estimated as follow s:

All of thermo-physical properties of air are determined at the overall bulk air temperature(Tb)from Eq.(6).

To assess the practical use,thermal performance of the enhanced tube is evaluated relatively to the smooth tube at an identical pumping power in the form of thermal enhancement factor(η)which can be expressed by

w here hpand hsare heat transfer coefficients for plain tube and inserted tube,respectively.

4.Results and Discussion

4.1.Validation of plain tube

In the present work,the heat transfer and pressure drop of plain tube in terms of Nussult number(Nu)and friction factor(f)are,respectively,verified with the Dittus–Boelter and Blasius correlations[28]respectively for,Nu and f as given in Eqs.(11)and(12):Dittus–Boelter correlation:

Blasius correlation:

Fig.3(a)and(b)show s that the measured data are in good agreement with the corresponding correlations.The average deviation of the measured is about 5%for Nu and 6%for f.

4.2.Effect of twisted-tape arrangements

The variation of Nu and Nusselt number ratio,Nu/Nu0,with Re for the tube with multiple twisted-tape inserts is depicted in Fig.4(a)and(b),respectively.It is visible that all the twisted-tapes yield considerable heat transfer with a similar trend in comparison with plain tube.For single twisted-tape,the heat transfer rate increases with the decreasing twist ratio as already mentioned in the literature.Due to the increase of swirl intensity imparted to the flow at tube wall,heat transfer rate of the inserted tube is found to be considerably higher than that of plain tube with no insert.This can be attributed to stronger swirl enhancing turbulence intensity,leading to higher convection heat transfer than axial flow in plain tube.Thus,the higher vortex flow,the greater Nusselt number becomes.In scrutiny of Fig.4,Nu obtained from the combined RTand LTtapes(called counter-twisted tapes)seems to be higher than that from the RT–RTcombination,or the LT–LTones(called co-twisted tapes).For double twisted-tapes,the counter-twisted tapes(2T2)perform better than the co-twisted tapes(2T1)and also the triple counter-twisted tapes(3T2)yield higher Nu than the triple co-twisted ones(3T1).In quadruple twisted tapes,it is noted that the four counter-swirl pairs,4T4 and four co-swirl pairs,4T1 provide,respectively,the highest and low est heat transfer rate.4T3 consisting of two counter-swirl pairs performs better than the 4T2 comprising the counter-swirl and co-swirl pairs.In addition,Nu obtained from wire coil insert taken from Ref.[3]is included in Fig.4 for comparison.The wire coil insert provides higher heat transfer rate than the twisted tape insert at all cases.

The Nusselt number ratio(Nu/Nu0)defined as a ratio of augmented Nusselt number to that of plain tube show s a decreasing trend with the rise of Re as can be seen in Fig.4(b).Under the present measured data,the Nu/Nu0values for 4T4,4T3,4T2,4T1,3T2,3T1,2T2,2T1 and 1T1 tapes are in the range of 1.15 to 2.12 depending on Re,while the wire coil yields the Nu/Nu0value ranging from 2.08–2.28,higher than the twisted tapes[Fig.4(b)].

Fig.3.Verification of(a)Nu and(b)f for plain tube.----correlations;□exp.data.

The influence of different twisted-tape number and arrangements on f and f/f0characteristics against Re is depicted in Fig.5(a)and(b),respectively.It is observed that the application of twisted tapes gives rise to considerably higher f than that of plain tube with no insert.The higher friction loss mainly comes from the increased surface area and higher swirl intensity.The f values for double and triple counter-twisted tapes(2T2 and 3T2)are seen to be higher than those for double and triple co-twisted ones(2T1 and 3T1)and are around 5%and 50%above that for single twisted-tape.For quadruple twisted-tapes,the highest and lowest f values are,respectively,found for four counter-twisted(4T4)and four co-twisted(4T1)tapes while the f of two counter-twisted tapes(4T3)is slightly larger than that of two co-twisted ones(4T2).This indicates a significant effect of multiple twisted-tape arrangements on thermal performance.Also,the f value of the wire coil taken from Ref.[3]is substantially higher than that of twisted-tapes.In Fig.5(b),the f/f0for twisted tapes is in the range of 1.94 to 4.06 while that for wire coil is ranging from 7.38 to 7.85.

Fig.4.Variation of(a)Nu and(b)Nu/Nu0 with Re for various twisted tapes.■4T4;?4T3;4T2;□4T1;3T2;△3T1;2T2;◇2T1;1T1,y/w=4;○1T2,y/w=5;wire coil[3];plain tube.

4.3.Effect of twisted-tape number

Fig.6(a)and(b)displays an effect of the number of counter-twisted tapes on Nu/Nu0and f/f0,respectively.It is visible that Nu/Nu0increases with the increment of N,especially for N＞2,but with the reduction of Re.The mean Nu/Nu0and f/f0increases for N=4(4T4),3(3T2),2(2T2)and 1(1T1)are,respectively,about 87%,68%,35%and 27%;and 278%,208%,111%and 107%above the plain tube.This shows the advantage of multiple counter-twisted tapes over the plain tube.

The axial wall temperature distribution along the tube at each thermocouple location for a four counter-twisted tape insert(4T4)is depicted in Fig.7.The wall temperature variation is found to gradually increase downstream of the inlet until x/D≈5–6 and then increased slightly to the exit.How ever,in the region near the exit,the wall temperature drops slightly from location x/D≈18 to the outlet,similar trend as found in Ref.[2]if it is presented in the form of local Nux.The wall temperature and the temperature difference between the inlet and outlet of air at low Re are seen to be higher than those at high Re.

Fig.5.Variation of(a)f and(b)f/f0 with Re for various twisted tapes.■4T4;?4T3;4T2;□4T1;3T2;△3T1;2T2;1T1,y/w=4;◇2T1;○1T2,y/w=5;wire coil[3];plain tube.

In addition,the empirical correlations developed by relating the Re and N together are compared with experimental data within±3.5%and±4.8%for Nusselt number and friction factor,respectively.Correlations for multiple counter-twisted tapes are

4.4.Thermal performance assessment

Fig.6.Effect of N on(a)Nu/Nu0 and(b)f/f0.Re=5363;Re=11660;Re=17638;Re=23057.

Fig.8.Variation of η with Re.■ 4T4;? 4T3;4T2;□4T1;3T2;Δ 3T1;2T2;◇2T1;1T1,y/w=4;○1T2,y/w=5;wire coil[3].

Figs.8 and 9 present the variation of thermal enhancement factor(η)with Re and the number of counter-twisted tapes(N),respectively.It can be seen in the plots that the η values are generally above unity for all the inserted tubes,except for the single twisted(1T1)and double co-twisted tapes(2T1)at higher Re values.η shows downtrend as Re increases for all tapes tested.Counter-twisted tapes provide higher η than co-twisted tapes,and the maximum η of about 1.33 is found for quadruple counter-twisted tapes(4T4)at low er Re.It is noted that η tends to increase with the increment of N.To obtain higher η,the use of counter-twisted tapes at N＞2 is recommended.The application of counter-twisted tapes leads to the increase in mean η of about 20%,15%and 5%for N=4(4T4),3(3T2)and 2(2T2),respectively.In comparison with wire-coil inserts[3],the triple and quadruple counter twisted tapes perform much higher than the wire-coil.How ever,η of wire coil is better than using single twisted-tape alone as seen in Fig.8.The detail on experimental results of multiple counter-twisted tapes is presented in Table 1.

5.Conclusions

An experimental study has been conducted to examine heat transfer and flow friction characteristics in a circular tube inserted with single,double,triple and quadruple twisted tapes comprised of either co-or counter-twisted tapes for the turbulent regime(Re from about 5300 to 24000)under a uniform wall heat- flux.For using multiple twisted tapes,Nu is in the range of 1.15–2.12 times while f is 1.94–4.06 times that of the plain tube.Both Nu and f are seen to increase with the increment of N and with counter-twisted tape arrangements.η tends to decrease with the increase in Re but with decreasing N.The highest η is found to be about 1.33 for using quadruple counter-twisted tapes(4T4).

Fig.7.Axial wall temperature distribution along test tube.

Fig.9.Variation of η with N.Re=5363;Re=11660;Re=17638;Re=23057.

Table 1Experimental results of multiple counter-twisted tapes

Nomenclature

A heat transfer surface area,m2

Cpspecific heat of fluid,J·kg?1·K?1

D inner diameter of test tube,m

f friction factor

h heat transfer coefficient,W·m?2·K?1

k thermal conductivity of fluid,W·m?1·K?1

L length of the test section,m

m mass flow rate,kg·s?1

N number of twisted tapes

Nu Nusselt number(=hD/k)

ΔP pressure drop,Pa

Pr Prandtl number(=Cpμ/k)

Q heat transfer rate,W

Re Reynolds number(=UD/ν)

T temperature,K

~T mean temperature,K

U mean air velocity,m·s?1

w tape width,m

y pitch length of twisted tape(180°rotation),m

η thermal enhancement factor[=(Nu/Nu0)/(f/f0)1/3]

ρ fluid density,kg·m?3

ν kinematic viscosity,m2·s?1

Subscripts

a air

b bulk

conv convection

i inlet

o outlet

pp pumping power

p plain tube

s swirl flow generator

w wall

Appendix

In order to evaluate the reliability of the experimental facility,the uncertainties of experimental data were determined.The uncertainties of the friction factor and Nusselt number data can be expressed as follow s

Friction factor:

w here Δ(ΔP)/ΔP= Δh/h and ΔRe/Re=[(Δm/m)2+(ΔD/D)2]0.5

Nusselt number:

w here h=q″/(Tw?Tb)

w here q″=(1/πDLh)[mCp(To? Ti)].