oceeding of the th WSEAS Int. Conf. on HEAT TRASFER, THERMA EGIEERIG and EVIROMET, Elounda, Greece, Augut -, 00 (pp-8) Experimental tudy of the heat tranfer for a tube bundle in a tranverally flowing air BY HORIA ECUA, EEA IETA GHIZDEAU, GEORGE DARIE Faculty of Power Engineering Univerity Politehnica of Bucharet Splaiul Independentei, Sector 00 Bucharet, Romania Tel: +0 098 Fax: +0 099 ROMAIA E-mail: ngh@cne.pub.ro http://energ.pub.ro Abtract. - The implet form of cro flow heat exchanger may be regarded a a erie of identical heat tranfer urface in a tranvere tream that each ha an influence on, and i in turn influenced by it neighbour. Therefore, in order to obtain a prediction for the heat tranfer rate to or from a bundle of urface in cro flow it i uual to initially conider a ingle urface in iolation a a bai for correlation. In one of the mot common arrangement, heat i tranferred between a fluid flowing through a bundle of tube and another fluid flowing tranverely over the outide of the tube. The main goal of thi tudy i the experimental determination of the convective heat tranfer coefficient tranferred between a fluid, which i the air, flowing through a bundle of tube in a tranverally flowing air in a taggered arrangement and the comparing with the theoretical correlation. Keyword: heat exchanger, bundle of tube, taggered arrangement, convective heat tranfer coefficient. omenclature: Greek ymbol d h n u m w hydraulic diameter [m] coefficient uelt umber, coefficient u = d λ α h ν andtl umber, = a w d ynold umber, = h ν pitch [m] air velocity in the minim ection α convective heat tranfer coefficient [W. m -. K - ] ε correction coefficient function of the number of tube row croed by the fluid ν air cinematic vicoity [m. - ] Subcript ymbol urface row number longitudinal T tranveral
oceeding of the th WSEAS Int. Conf. on HEAT TRASFER, THERMA EGIEERIG and EVIROMET, Elounda, Greece, Augut -, 00 (pp-8). ITRODUCTIO The heat exchanger are thermal equipment preent in almot all indutrial ector, playing an eential role in many procee and ytem. Increaing the efficiency of thi equipment determine functioning condition and performance of technological aemblie. Heat tranfer to or from a bundle of tube in cro flow i relevant to numerou indutrial application uch a team generation in a boiler or air cooling in the coil of an air conditioner. The overall heat tranfer coefficient for a cro flow heat exchanger i made up of three component: the urface heat tranfer coefficient for the fluid flowing through the tube, the thermal conductivity and thickne of the tube material and the urface heat tranfer coefficient for the fluid flowing over the external urface of the tube. For a better dimenioning of thee device there i neceary a better evaluation of the convective heat tranfer coefficient []. The characteritic of the heat exchanger can be etablihed wither directly by experimental meaurement or by numerical imulation. The experimental meaurement are needed in order to develop new the new heat exchanger deign, and for the etablihment of the optimal operational parameter []. In the cae of a ga, which in our cae i air, flowing throughout a bundle of tube, the aement of the effective heat tranfer coefficient i very important eeing that in general there are the lowet convective heat tranfer coefficient, which influence the effective global heat tranfer coefficient. Making a piece of equipment a compact a poible for obtaining a heat tranfer rate a big a poible i the main concern in the reearch activity of heat exchanger.. EXPERIMETA DEVICE Our experimental device preented in figure conit in the air duct, which i vertically mounted gla reinforced platic duct with bell mouth intake at it upper end. The fan i mounted on an epoxy coated welded teel frame. Air duct i directly mounted on the frame and fan intake. Row Flow direction Tube may be removed and replaced by heated cylindrical element Fixed Figure. Experimental etup The active element i a thick cylinder electrically heated. The maximum temperature reached by the active element i 00 º C. Extreme end are inulated to reduce error due to wall effect. Integral thermocouple ene urface temperature. The clear platic plate with 7 fixed platic tube of mm nominal diameter arranged in a taggered pitch with the longitudinal pitch = 7. mm, and the tranvere pitch T = mm. ear the centre of each row i a dummy tube that may be removed and replaced with the active element. All electronic intrumentation and control i houed in a platic coated teel conole which conit of the digital electronic thermometer with 0. reolution, which indicate element urface temperature and, via a biaed witch, the duct air temperature and an analogue voltmeter indicating the voltage acro the active element heater. The maximum voltage i 70 V. The preure i meaured by duct mounted inclined manometer recording intake depreion one with the range from 0 to 70
oceeding of the th WSEAS Int. Conf. on HEAT TRASFER, THERMA EGIEERIG and EVIROMET, Elounda, Greece, Augut -, 00 (pp-8) mmh O and the econd one from 0 to 0 mmh O. On thi experimental device it can be made the following experiment: Steady tate determination of the mean urface heat tranfer coefficient for tube in the t, nd rd, th, th and th row of a cro flow heat exchanger Determination of the mean urface heat tranfer coefficient for cro flow heat exchanger with one to ix row Deduction of the relationhip between uelt, ynold and andtl number for each of the ix tube row. The experimental tudy wa realized with the device preented above. The active element wa introduced into the top open hole in the tube plate and plug the lead into the element conole. The other five remaining dummy tube are left in poition in the lower hole. The duct preure i connected of the lower manometer and the fan i et up to the lower poition which correpond with the fan cloed in order to obtain a low velocity The device i tarted up there i adjuted the heater control by increaing the voltage to give an indicated element urface of approximately 9 ºC. When table condition are occurred indicated by a contant active element urface temperature are recorded the active element urface temperature, the air temperature, the preure and the voltage. After that the iri damper of the fan exhaut i adjuted in order to increae the air velocity. By iri dumper of the fan exhaut variation the velocity range i etup between m/ and m/. In thi new cae the heater control i increaed to give approximately the original active element urface temperature. Again, when table there are recorded the ame variable. The procedure i repeated by increaing the air velocitie up to the maximum, which correpond of an iri damper fully open. After that the element i cooled and it i placed in the econd row hole and the dummy tube from thi hole i placed in the firt row hole. The whole experiment i repeated for a imilar range of air velocity. The entire procedure i repeated with the active element in the row,, and..theoretica CORREATIO Turbulent flow condition do not lend themelve to imple theoretical analyi and therefore alternative method are required in order to evaluate urface heat tranfer coefficient for general flow condition. One of thee method i to apply the principle of dynamic imilarity, which can be ummarized by writing []: (, ) u = f () Uing the dimenional analyzing the general relation () reult in the following relation: m n u = C () The tube poition within the bundle add further variable to the general turbulent flow equation () and therefore thi ha the form: m n u = C ε () The value of the contant C and m are function on geometric parameter and tube bundle arrangement. The value of contant n are fluctuating through 0. 0. and 0.. ε repreent a correction coefficient function of the poition of tube row in the bundle []. Alternatively, there are preented the mot ued correlation applicable decribed in the table.
oceeding of the th WSEAS Int. Conf. on HEAT TRASFER, THERMA EGIEERIG and EVIROMET, Elounda, Greece, Augut -, 00 (pp-8) Table The mot ued correlation for u in a taggered arrangement Correlation Iachenko [] u = 0. 0 0. T. Kay [] 0. 0. u= 0. ε Miheev [] u = 0 0. 0.. 0. ε ε Applicability condition 0 < < 0 0,7< < 00 0,< 0 0,7< < 00. > 0 <. Obervation =, ε =, ε, ε = ε = 0. = ε = 0.7 =. ε = 0.8 ε = 0.7 ε = 0.8 ε = 0.89 ε = 0.9 ε = 0.9 ε = 0.97 ε = 0.98 ε = 0.99 7 ε =.0. 0 =, ε =, ε, ε 8 =. 9 = ε = 0. = ε = 0.7 Zhukauka [] 0. 0. 0 T 0. 0.. 0 < < 0 u = ε T <. Grimion [] 0 < < 0 m u = C ε 0,7 ε = 0. ε = 0.7 ε = 0.8 ε = 0.89 ε = 0.9 ε 7 = 0.9 ε = 0.97 ε = 0.98 ε = 0.99.. 0. T D D T C, m= f, D D In our cae C=0. and m =0.. 0 ε = 0.8 ε = 0.7 ε = 0.8 ε = 0.89 ε = 0.9 ε = 0.9 ε = 0.97 ε = 0.98 ε = 0.99. 7 8 9. EXPERIMETA STUDY The determination of the correlation for the mean convective heat tranfer coefficient for the tube forming a cro flow heat exchanger i carried out experimentally uing the device preented above. The meaurement were made for each row of the tube bundle, for different air velocitie, which implied different ynold number. the reult preented in the figure for a tube bundle in a taggered arrangement. In the figure there i preented the variation of u with for each row from to. Uing the correlation decribed in the table and the experimental data there are obtained
oceeding of the th WSEAS Int. Conf. on HEAT TRASFER, THERMA EGIEERIG and EVIROMET, Elounda, Greece, Augut -, 00 (pp-8) u 0 0 00 0000 000 0000 000 0000 Row Row Row Row Row Row u 0 0 00 0000 000 0000 000 0000 Row Row Row Row Iacenko Kay Miheev Zukauka Grimion Figure Experimental data for each row In figure and there are preented the comparion of the variation of u with through the experimental data and the mot ued correlation: Iacenko, Kay, Miheev, Zukauka, and Grimon for the t row and nd row. u 0 00 90 70 0000 000 0000 000 0000 experimental Iacenko Kay Miheev Zukauka Grimion Figure Comparion through experimental data and theoretical correlation for the,,, and row For the model created there i obtained the correlation between u and and determined the value of the contant C and m. Therefore, according to figure, on the linear regreion of the experimental data for the row to there wa obtained ame value 0. for the coefficient C and m. Thi value might be conidered with precaution due a few number of experimental point, but however it i approached to the value exiting in literature for thi range of ynold. Figure Comparion through experimental data and theoretical correlation for the t row y = 0,x 0, u 0 0 0 00 90 70 experimental Iacenko Kay Miheev Zukauka Grimion u/ 0. 0 Row - 0 00 0000 000 0000 000 0000 0000 000 0000 000 0000 Figure Comparion through experimental data and theoretical correlation for the nd row. Starting with the rd row there i oberved that the experimental data are practically uperpoed. Therefore there wa born the idea of adopting a model, which i auming the unification of the heat tranfer from the rd row above, imilar with Iachenko and Miheev model. The reult are yntheized in the figure. Figure Determination of the value of the contant C and m. For the row and, the correction coefficient obtained have the value: ε = 0. and repectively ε = 0.79, according to the correlation obtained and preented in figure.. COCUSIOS In the experiment performed there i obtained the variation between u and for from 0000 to 0000, our range of tudy. For thi tudy there wa oberved that
oceeding of the th WSEAS Int. Conf. on HEAT TRASFER, THERMA EGIEERIG and EVIROMET, Elounda, Greece, Augut -, 00 (pp-8) the minimum error for our regreion i obtained for a coefficient of andtl equal to n = 0.. Comparing the experimental data for uelt number with the data obtained form figure, reult a range of relative error of ( % % ) for the whole experiment achieved. The tandard relative error, related to the theoretical correlation, for thi type of experiment are preented in table for the row to. For the row to the correcting factor for ε have a imilar value with the theoretical data preented in table. The uelt number for the row to obtained experimentally, are not influenced anymore on the corrective factor variable ε of each row, therefore there wa adjuted a common relationhip for thee row. Table Theoretical relative error Author lative error min. max. Iacenko -% 9% Kay -% % Miheev -% % Zukauka -0% % Grimion -0% -% A a concluion, chooing the bet correlation for a flowing throughout a bundle of tube, hould be made carefully, due to the error that could be obtained, error that hould be avoid a much a poible. And upward the error hould be minimum due to the etimation of the global tranfer heat coefficient for the whole device. Moreover, for example, for an air heater, uing different correlation for convective heat coefficient i influenced the heat urface. In thi cae, the burned gae are paed through the tube, and the air i paed tranverally over outide the tube. Uing the correlation decribed above, there are obtained the following relative error for the heat urface (ee table ). Table lative error for heat urface for different correlation (ference Miheev correlation) Author Global heat urface relative error Iacenko.% Kay.% Zukauka 0.% Grimion.0% Therefore for obtaining the optimum olution the correlation hould be choen very carefully. Thi cae i obtained when the air convective coefficient i greater than the ga convective coefficient. If the air convective coefficient i maller than the ga convective coefficient than the relative error are maller. All thee reult are the firt tep in the heat exchanger reearch. Thi tudy may be continued with the reearch of other geometrie, other arrangement, other fluid. Moreover a very intereting reearch i the imulation in FUET and CFX code, for all thee new reearch direction. ference: []. Iachenco V. P., Oipova V. A., Sukomel A. S. Heat tranfer. Mocow. Mir Publiher. 977. []. Kay W. Convective heat and ma tranfer. ew York. McGraw-Hill, 99. []. Badea A. Bazele tranferului de căldură şi maă. Bucharet. Romania Academy Publiher. 00. []. Zukauka A. - High-performance ingle-phae heat exchanger. ew York. Hemiphere Publihing Corporation. 989. []. Incropera F. P., DeWitt D. P. Fundamental of heat and ma tranfer (fourth edition). U.S.A. John Wiley & Son. 99. []. A. Horvat, M. ekovar, B.Mavko Comparion of heat tranfer condition in tube bundle cro-flow for different tube hape. Int. J Heat Ma Tranfer 9 (00) 07-08.