A computer based iterative solution for accurate estimation of heat transfer coefficients in a helical tube heat exchanger

A computer based iterative solution for accurate estimation of heat transfer coefficients in a helical tube heat exchanger
of 4
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A computer based iterative solution for accurate estimation of heat transfer coefficients in a helical tube heat exchanger P.K. Sahoo  a , Md.I.A. Ansari  b , A.K. Datta  b,* a Department of Food Engineering, Faculty of Agricultural Engineering, Bidhan Chandra Krishi Viswavidyalaya, Mohanpur 741 252, Nadia,West Bengal, India b Department of Agricultural and Food Engineering, Indian Institute of Technology, Kharagpur, West Bengal 721 302, India Received 19 February 2002; accepted 09 September 2002 Abstract An iterative technique is developed and reported for accurate estimation of heat transfer coefficients in a helical triple tube heatexchanger. Based on the experimental temperature rise of whole milk in a helical triple tube heat exchanger, accurate values of filmheat transfer coefficients and overall heat transfer coefficients based on the outside area of the innermost tube and the inside area of the middle annulus are obtained from first principles. Three different flow rates of milk were used giving rise to three Reynoldsnumbers. The procedure described can be applied to any heat exchanger with minor modifications, if necessary.   2003 Elsevier Science Ltd. All rights reserved. Keywords:  Heat transfer coefficient; Heat exchanger; Iterative solution; Helical tube 1. Introduction Accurate estimation of heat transfer coefficients inheat exchangers is difficult, particularly in tubular heatexchangers where an annulus is used to heat or cool theworking fluid. Since both the inside and outside surfacesare responsible for heat transfer, it is important that thefilm coefficient in the working fluid is accurately esti-mated. In turn the overall heat transfer coefficientsbased on the inside area of the middle tube and theoutside area of the innermost tube should be obtained.In this article an iterative computer based solution isreported for accurate estimation of such heat transfercoefficients. Experimental data on a concentric helicaltriple tube heat exchanger (Fig. 1) is used to validate theprocedure. The middle tube of the heat exchanger wasused to convey the working fluid (i.e. milk). The inner-most and outermost tubes were used to carry the heatingmedium (i.e. steam). The temperature rises of milk atthree different flow rates were obtained. The iterationwas based on equations obtained from first principlesand is described in Section 2. 2. Theory The condensate film heat transfer coefficients wereobtained from the following equations (Kern, 1984): h c1  ¼ 0 : 5754  k  3c q 2c  g  kl c  D i1 ð T  s  T  i1wall Þ   0 : 25 ð 1 Þ h c2  ¼ 0 : 725  k  3c q 2c  g  kl c  D o2 ð T  s  T  o2wall Þ   0 : 25 ð 2 Þ The condensate film temperature ( T  f  ) was obtained fromthe following equations (Cuevas & Cheryan, 1982): T  f1  ¼ T  s  0 : 75 ð T  s  T  i1wall Þ ð 3 Þ T  f2  ¼ T  s  0 : 75 ð T  s  T  o2wall Þ ð 4 Þ Initial estimation of the film heat transfer coefficient onthe milk side was carried out using the following equa-tion (Chopey & Hicks, 1984; Geankoplis, 1997; Zuritz,1990): h m  ¼  k  m  D eq   0 : 027  N  0 : 8  Re  N  0 : 33  Pr   D i2  D o1   0 : 53 1   þ 3 : 5  D eq  D c   ð 5 Þ T  i1wall  and  T  o2wall  were obtained from the followingequations based on the heat flux equivalence concept(Cuevas & Cheryan, 1982): Journal of Food Engineering 58 (2003) 211– * Corresponding author. E-mail addresses: (P.K. Sahoo), (Md.I.A. Ansari), (A.K. Datta).0260-8774/03/$ - see front matter    2003 Elsevier Science Ltd. All rights reserved.doi:10.1016/S0260-8774(02)00370-9  T  s  T  i1wall1 h c1  D i1 ¼  T  i1wall  T  bmln  D o1  D i1   2 k  ss þ  1 h m  D o1 ð 6 Þ T  s  T  o2wall1 h c2  D o2 ¼  T  o2wall  T  bmln  D o2  D i2   2 k  ss þ  1 h m  D i2 ð 7 Þ Overall heat transfer coefficients were obtained from thefollowing equations based on conduction and convec-tion in cylindrical coordinates: U  o1  ¼  1 h m 0@  þ  D o1 ln  D o1  D i1   2 k  ss þ  D o1 h c1  D i1 1A  1 ð 8 Þ U  i2  ¼  1 h m 0@  þ  D i2 ln  D o2  D i2   2 k  ss þ  D i2 h c2  D o2 1A  1 ð 9 Þ Introducing q 1  ¼ U  o1  A o1 D T  lm  ð 10 Þ and q 2  ¼ U  i2  A i2 D T  lm  ð 11 Þ _ mm m C  m D T  m  ¼ q 1 þ q 2  ð 12 Þ 3. Methodology The dimensions of the tubes in the concentric helicaltriple tube heat exchanger are shown in Table 1. Wholemilk was preheated to 93   C in steam jacketed pan andsupplied to the inlet of heat exchanger by means of acentrifugal pump. Milk was raised to various tempera-tures inside the heat exchanger by steam supplied from a‘‘VASPA’’ electrode steam boiler. The various outlettemperatures for different flow rates are given in Table 2.The developed iterative method consists of the followingsteps:1. Tube diameters, coil diameter, flow rate of milk,length of helical triple tube heat exchanger, tempera-ture of steam, inlet and outlet temperatures of milkare the inputs.2. Milk side film heat transfer coefficient is estimated us-ing Eq. (5).3. Wall temperatures ( T  i1wall  and  T  o2wall ) are initializedby assuming  T  i1wall  <  T  s  and  T  o2wall  <  T  s .4. Condensate film temperatures are estimated usingEqs. (3) and (4).5. Condensate film heat transfer coefficients are ob-tained using Eqs. (1) and (2). Nomenclature  A  heat transfer area, m 2 C   specific heat, kJ/kgK  D  tube diameter, m  D c  coil diameter, m  g   acceleration due to gravity, 9.81 m/s 2 h  film heat transfer coefficient, W/m 2 K k   thermal conductivity, W/mK  L  length of tube, m _ mm  mass flow rate, kg/s  N   dimensionless number T   absolute temperature, K U   overall heat transfer coefficient, W/m 2 K Greek symbols D  difference k  latent heat of condensation, kJ/kg l  dynamic viscosity, kg/ms q  density, kg/m 3 Subscripts b bulk or average between two heat transfersurfacec condensate filmh heatingeq equivalenti insidem milko outsides steamss stainless steellm log mean Re  Reynolds number Pr  Prandtl number1 tube of smallest diameter2 tube of intermediate diameter Fig. 1. Isometric view of helical tube UHT milk sterilizer.212  P.K. Sahoo et al. / Journal of Food Engineering 58 (2003) 211–214  6.  T  i1wall  and  T  o2wall  are obtained using Eqs. (6) and (7).7. Steps (3)–(6) are repeated until  T  i1wall  and  T  o2wall assumed in step (3) matches with those obtained instep (6).8. Overall heat transfer coefficients ( U  o1  and  U  i2 ) arecalculated using Eqs. (8) and (9).9.  q 1  and  q 2  are obtained using Eqs. (10) and (11).10. Obtain ratio  r  ¼ q 2 q 1 .11. Substitute experimental data and  q 2  ¼ rq 1  in Eq. (12).12. Maintaining ‘‘ r  ’’ obtain new  q 2 .13. New  U  o1  and  U  i2  are calculated from Eqs. (10) and(11).14. New  h m  is calculated by using new values of   U  o1  and U  i2  in Eqs. (8) and (9) and obtaining an average of thetwo values.15. Steps (8)–(13) are repeated until the two sets of over-all heat transfer coefficients match.16. Replace the value of milk side film heat transfer coef-ficient ( h m ) in step (2) with the value obtained in step(14).17. Repeat step (7).18. Repeat step (15).19. Stop when  h m  obtained in step (14) in the last itera-tion is within   0.01% of the previous iteration. 4. Results and discussion Two sets of values of milk side film heat transfercoefficient ( h m ) and overall heat transfer coefficients ( U  o1 and  U  i2 ) are presented in Table 3. The ‘‘calculated val-ues’’ refer to those obtained after completing all thesteps involved in the iteration whereas the   predictedvalues   refer to the values obtained on the basis of Eq. (5)and iteration completed up to step (13) only. The reasonfor such comparison is to obtain the effect of the semi-empirical Eq. (5) on the fluid values of overall heattransfer coefficients. Fig. 2 compares the two sets of values of milk side film heat transfer coefficients forvarious Reynolds numbers. Fig. 3 plots the ‘‘predicted’’and ‘‘calculated’’ values of overall heat transfer coeffi-cients based on outside area of the innermost tube asaffected by Reynolds numbers. Fig. 4 is similar except Fig. 2. Variation of milk side heat transfer coefficient with Reynoldsnumber.Fig. 3. Variation of   U  o1  with Reynolds number.Table 1Dimensions of tubes in helical triple tube heat exchangerInnermost tube diameters (mm) Middle tube diameters (mm) Outermost tube diameters (mm) Length (m)Inside Outside Inside Outside Inside Outside6.00 8.50 12.7 15.0 22.23 25.4 2.28Table 2Milk flow rates, Reynolds numbers and outlet temperaturesFlow rate (l/h) Reynolds number Outlet temperature (  C)135 5159 142153 5791 140176 6587 138Table 3Calculated and predicted values of heat transfer coefficientsReynoldsnumberCalculated film heattransfer coefficient( h m ) (W/m 2 K)Predicted film heattransfer coefficient( h m ) (W/m 2 K)Calculated  U  o1 (W/m 2 K)Predicted  U  o1 (W/m 2 K)Calculated  U  i2 (W/m 2 K)Predicted  U  i2 (W/m 2 K)5159 1508 1516 1226 1102 1318 14005791 1587 1596 1275 1140 1376 14666587 1701 1713 1345 1196 1459 1558 P.K. Sahoo et al. / Journal of Food Engineering 58 (2003) 211–214  213  that it plots the overall heat transfer coefficient valuesbased on the inside area of the middle tube. The mag-nitudes of difference between the two sets vary from0.55% to 0.70% for milk side heat transfer coefficient.The differences between  U  o1  values vary from 10% to11%. The same difference between  U  i2  may vary from6.00% to 7.00%. The source of error in the experimentaldata is in the measurement of volume flow rate of liquidin the heat exchanger. Since every liter measurement hasa   10 ml error we can say that there is a   1% error inthe reported Reynolds  s numbers. This shows up as a  0.8% error in the values of film heat transfer coeffi-cients  h m . The corresponding error values for  U  o1  and U  i2  are   0.72% and   0.77% respectively. So for filmheat transfer coefficients the experimental error is com-parable to the errors between predicted and calculatedvalues. On the other hand the experimental error isnegligible compared to the spread between calculatedand predicted values of   U  o1  and  U  i2 . The method dem-onstrates how accurate estimation of heat transfer co-efficients is possible when two heat transfer surfaces areinvolved. This method is easily adaptable to any con-centric heat exchanger involving two heat transfer sur-faces. The input to the programme must be altered tosuit other tubular heat exchangers. Appropriate heattransfer equation also has to be substituted for Eq. (5). References Chopey, N. P., & Hicks, T. G. (1984).  Handbook of chemical engineering calculation  (p. 33). New York: McGraw-Hill Bookcompany (Chapter 7).Cuevas, R., & Cheryan, M. (1982). Heat transfer in a vertical liquidfull scraped surface heat exchanger: application of the penetrationtheory and Wilson plots method.  Journal of Food ProcessEngineering, 5 , 1–21.Geankoplis, C. J. (1997).  Transport processes and unit operations  (3rded., p. 239). New Delhi: Prentice Hall of India Pvt. Ltd.Kern, D. Q. (1984).  Process heat transfer  (pp. 263–269). New York:McGraw-Hill International.Zuritz, C. A. (1990). On the design of triple concentric tube heatexchangers.  Journal of Food Process Engineering, 12 , 113–130.Fig. 4. Variation of   U  i2  with Reynolds number.214  P.K. Sahoo et al. / Journal of Food Engineering 58 (2003) 211–214

Caso La Minerva

Mar 4, 2018
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks