HEAT 008, Fith International Conerence on Transport henomena In Multiphase Systems June 30 - July 3, 008, Bialystok, oland A Numerical Study o an Evaporator Coil or a Rerigeration Secondary Loop with CO Z. Aidoun 1, M. Ouzzane 1 1 CTEC-Varennes, Natural Resources Canada 1615, Boulevard Lionel Boulet, Varennes, Québec, H3X 1S6, Canada, zaidoun@nrcan.gc.ca and mohamed.ouzzane@nrcan.gc.ca Abstract Environmental concerns have dictated the identiication and the development o substitutes or the current synthetic rerigerants replacement. Carbon dioxide is a natural alternative suitable or medium and low temperatures. Heat transer between the rerigerant and the rerigerated medium occurs in the evaporator, whose design must be adapted to take advantage o the avorable properties o carbon dioxide. This paper presents a mathematical model to study in detail a counter-current type air/co coil with corrugated ins. The model applies the conservation equations or mass, momentum and enthalpy on small volume elements. The solution procedure o these equations is based on the Forward Marching Technique. The computed results allow or detailed inormation on the air and rerigerant states in the coil. Local results can be integrated to obtain global values or engineering purposes. It is possible to qualiy tube rows by determining their individual capacities in terms o their location in the coil. Experiments have been perormed on an experimental acility in CTEC-V Laboratories. Results rom this installation as well as those collected rom the open literature have been used to validate the developed model. Agreement between the predictions and experimental values is satisactory. Calculations perormed on a typical 3.5 kw capacity rerigerated cabinet coil at -4 C, highlight some advantages o CO : - small CO side pressure drop and small corresponding temperature variation - nearly uniorm temperature and humidity distributions on the air side. - possibility o using long circuits or simple coil geometry and impacting positively on rerigerant distribution and the cost o the coils. INTRODUCTION As a natural luid oering environmental advantages, carbon dioxide has thermo physical and transport properties that places it among the preerred alternatives or synthetic rerigerants such as HCFCs and HFCs. In order to ully exploit the advantages related to the use o carbon dioxide, several researchers over the last ew years have devoted time and eort to clariy the issue. Among the pioneering works were those o Lorentzen (1993 and 1995). Since then research was perormed on systems, equipment and heat transer processes involving this rerigerant. In the case o supermarket applications, particular interest is directed towards the use o secondary loops rerigeration. Such an approach based on CO with phase change is a potentially interesting alternative to dry expansion or sensible heat transer, since pressures are within the range o traditional rerigeration applications at moderate temperatures. In such cases, the combined CO latent heat due to phase change with its excellent thermophysical properties, signiicantly reduce the total charge and rerigerant low rates, as it is clearly outlined by earson (1993) and Inlow et al. (1996) respectively, just to mention these two papers. However in order to be competitive with existing direct evaporation systems, secondary loops must be designed such that ineiciencies related to extra heat exchangers and temperature gradients are minimized. Other developpements continue to be proposed. For example Sawalha et al. (005) have studied the case o a rerigeration system that simulated real operating conditions in a supermarket using CO as rerigerant. Dispenza et al. (005) perormed a easibility study on a hypermarket in Sicily (south o Italy), in which they recommended two dierent conigurations: the irst coniguration uses carbon dioxide in a three-stage supercritical cycle; the second coniguration combines the use o propane and CO in cascade. Both options appear to be suitable or supermarket rerigeration, based on the TEWI criterion which is comparable or both cases. However the technology is considered to be still at an R-D stage. Modeling work on evaporator coils such as is presented here has also received a great deal o attention rom researchers because coils are the main heat transer component in a rerigeration set up at the cold end. Moreover most o this work is also applicable or rerigerants other than CO. In this respect, Liang et al. (1999) have used rerigerant R134A in their evaporator coil. However the computation approach is also applicable to CO. Model validation was based on global perormance parameters such as rerigerating capacity. More recently, H. Jiang et al. (006) have modeled their heat exchanger making provision or the lexibility in the treatment o several rerigerant circuits. These
models use a tube by tube approach in their computations. The mathematical model proposed in this paper was developed to study in detail a counter-current type air/co coil with corrugated ins. The model applies the conservation equations or mass, momentum and enthalpy to a coil divided according to an incremental approach into small volume elements. In the process appropriate CO correlations or heat transer and pressure drop are used. The solution procedure o these equations is based on the Forward Marching Technique. The results obtained by means o these computations allow or detailed inormation on the luids states in the coil. On the air side, temperature, relative humidity and external pressure drop are among the important parameters that can be obtained. On the rerigerant side lowing in the tube, saturation temperature, and internal pressure drop and vapor quality along the coil can be tracked. Local results can be integrated to obtain global values or engineering purposes. It is possible to qualiy tube rows by determining their individual capacities as a unction o their location in the coil. As an example, calculations have been perormed on an evaporator coil typically used in rerigerated cabinets with three doors. The coil capacity is approximately 3.5 kw and the air inlet temperature is -4 C. The results obtained on the basis o this case have been used to highlight the main advantages o CO as a rerigerant. Conservation equations are applied on an element o control volume as represented in Fig.1-b. Q.. i = m a ( Ha i Ha i+ 1 ) = m co ( Hcoi+ 1 Hcoi ) (1) Ugi Qi = () Ugi. π.d ou. ΔLi. ΔTlmtd 1 Dou A ou Ln A ou D 1 in + + hin iain η hou i πλt. ΔL i = (3) ΔT lmtd is the logarithmic temperature dierence and Ug i is the overall heat transer coeicient. Air side pressure losses and the heat transer coeicient have been respectively calculated by using the Colburn and riction coeicients available rom Wang et al. (00). The Colburn equations or N rows are: 0.69 0. 737 ( tanθ) Nrang J J 0.0646. Re 1 = Dc.GF 1. (4) THEORETICAL MODEL Important parameters including rerigerant properties, geometry, coniguration layouts and other operating conditions signiicantly inluence operation and perormance. It is the purpose o this contribution to deal with the development o a mathematical model or design and analysis that will be used to harness these parameters or a rational design o air-co wavy in evaporator coils which are standard in rerigeration. Rerigerant CO lows inside horizontal tubes while air lows across the coil and over the ins, on the outside o the tubes. Flow o rerigerant through its progression in the coil can be distributed in sub cooled, two-phase and superheated regions along the tube and tube lengths are allocated according to the desired perormance and operating conditions, as detailed in Aidoun et al. (004). The resulting computer programs are designed to tackle staggered tube situations, accounting or co-current or counter-current lows and updating or pressure drop related temperature and rerigerant properties variations along the low. In the present work, this tool is irst used to perorm simulations o the eect o a staggered tube coil arrangement in counter low with phase change, on perormance. It is then applied to examine the parametric distributions under dierent operating conditions. erormance is expressed in terms o coil capacity, internal and external pressure drops, temperature distributions (air, wall and rerigerant). The model relies on the resolution o the conservation equations or mass, momentum and enthalpy, with the main assumptions stated below: - steady state low conditions; - one-dimensional low or CO inside tubes and air across the coil; - negligible thermal losses to the environment; - uniorm temperature and air low beore each pass; - no condensation and no rost accumulation. Re Dc is the Reynolds number based on the collar diameter. Fig.1-a: Coil coniguration
Fig. 1-b: Volume element J 1.03 0.43 Dc Fs 1 GF1 =. D h t D c 0.4 0.56 J1 0.0545 0.0538tanθ 0.30Nrang.GF. tanθ GF3 (5) = (6) 1.3 0.379 1.35 Fs 1 1 GF = 1 t D (7) h 0.166 1.08tanθ J = 1.9.GF 3 (8) Nrang 1.77 9.43tanθ 0.9 1.43tanθ 0.174ln( 0.5Nrang) 1 Dc Fs = t D h (9) t θ is the deviation angle o the in corrugation relative to the horizontal, D c and D h are respectively the external and hydraulic diameters. F s is the in spacing. l and t are respectively the longitudinal and transverse spacings o the tubes. Re Dc is the Reynolds number relative to the external diameter. The riction actor is deined as: 3 4 0.383 0.47 Fs 1 D c 1 l = 0.8 Re Dc.( tan θ) 1 D c D (10) h t Where: 0.47 0.049 + tan θ.( tan θ) rang 0.37 1 = 0.141.GF4 N (11) The heat transer coeicient calculation or CO with phase change has been calculated by using the correlations developed by Bennet-Chen and modiied by Hwang et al. (1997). These are based on the superposition principle which consists in assuming that the heat exchange coeicient is equal to the sum o two exchange coeicients respectively representing nucleate boiling heat transer and convective heat transer. h = h nb + h bc (16) Correlations based on the homogeneous model have been selected or the calculation o linear pressure losses, Rohsenow et al. (1998). Δ v tp(i) ) G (17) Δ L i =. ΔLi.v tp(i+ 1) + (v tp(i+ 1) Din ressure losses in bends have been determined by using the correlation due to Geary (1975). s in g L ben G.X Δ s = s (18) D. ρ d 6 in 0.5 g 803.5 10.Re = (19) 1.5 exp(0.15 C / D )X The properties o air above the dew point temperature have been determined rom the conventional psychometric relations, summarized in the ASHRAE Fundamentals Handbook, ASHRAE (1993). Cpm = (0) Cpa + ωcpv And the corresponding actors are: In the vicinity o the wall, the ollowing correlation is used: 0.051 0.35 0.449 tan θ F = s l t 4 1 t Dh GF (1) C pw H = T w w H T ev ev 0.093 0. 013 ( ln( Re Dc )) Nrang = 0.56 (13) (1) 3 0.06 0.03 1 = 0.30 Re Dc. (14) t 4 = 0.306 + 3.63 tan θ (15) H w and H ev are air enthalpies based on the wall and on the CO evaporation temperatures respectively. When the wall temperature is below the dew point and assuming that the air is saturated at the wall temperature condition, use o the ollowing equations is made: Cpm = () Cpa + ωm. Cpv and 3
ω + ωsat ω m = (3) 4 Expt. 1495.9-35.5 44.4 -.7-4.9 Model 1871.1 1798.4 0.0 107.7-4.9-8.0 Expt. 1640.0-18.3 97.0-4.5-7.8 SOLUTION ROCEDURE AND VALIDATION The modular structure o this model is such that it can handle dierent coil arrangements. The tubes are tagged in terms o the coordinate parameters Nligne (vertical position) and Nrow, (horizontal position). Once the coniguration, the tube length and other geometric characteristics are selected, the program sets the overall pattern with the number o passes over which computations will be perormed. The solution procedure adopted is based on an incremental approach which consists in resolving the equations applied on each volume element in the inverse direction o the CO low path. Iterations are perormed up to the point where inlet conditions are attained. During the computations the return bends are automatically sensed and the cumulative count o the total pressure losses is perormed. Thermodynamic properties o CO and air are respectively determined by using NIST- REFRO subroutines, NIST (1998) and conventional psychrometric relations, ASHRAE (1993). The experimental set up o the CTEC-Varennes has been used or validation. It is represented in Fig.. Evaporating carbon dioxide is the working luid in the loop o interest (L1), which includes a CO -air coil with aluminium wavy ins and copper tubes. The loop is well instrumented or the purposes o heat and mass transer balances and luid low. It is located in a closed, well insulated room with two compartments corresponding to the inlet and the outlet o the coil: air lows rom one compartment to the other through a duct enclosing the coil. Air circulation is maintained by a blower. The coniguration employed in this particular instance is similar to that represented on Fig.1-a, arranged in one circuit. Other characteristic details o the installation, measurements uncertainties and operational procedure can be ound in Ouzzane et al. (to appear in 008). The computed results have then been compared with those obtained by experiments perormed on the above-mentioned test bench. 1 Table 1 Comparison o results rom measurements and computations Case Capacity (W) X% Δ (ka) Te ( o C) Air CO CO CO Air CO Model 4438.3 4530.6 58.0 195. -3.0-9.9 Expt. 4083.3-51.4 171.6 -.4-9.6 Model 4615.9 4673.3 97.0 1.1-3.8-9.7 Expt. 4556.7-90.0 111.4-3.7-9. 3 Model 1557. 1550.0 36.5 38.6 -.8-4.8 4 These are shown in Table 1, in terms o capacity, quality, internal pressure drop and temperatures on the air side and the rerigerant side respectively. The evaporator coil experimented upon has the ollowing characteristics : - rontal section o 0.61m 0.3m or air low and a depth o 0.m. - the tubes are inned with corrugated aluminium ins in the number o 157 ins per meter. - the coil is 8 rows deep and 10 rows high. The operating conditions are: - air inlet temperatures between -1 C and -10 C. For CO temperatures vary between -7 C and -3 C. - air velocities in the range o 1m/s to 3 m/s - CO mass low rates in the range o 0.00 kg/s to 0.100 kg/s. The heat balance that links the air side and the rerigerant side is very good. The agreement between the model predictions and the experimental data is satisactory. In the worst case the predictions are within 1 % o the experimental values. The temperatures on the air side and the rerigerant side are particularly well predicted. In any case the predictions are all within the uncertainty range o the used correlations as stated by their authors. RESULTS The prediction results obtained by using the developed calculation tool reer to a practical case selected or this study and are based on an existing evaporator coil commonly used in three-door reezing cabinets. It is o a staggered tube coniguration and counter-current low, as depicted in Fig. 1-a. Rerigerant CO in phase change vaporizes while lowing inside tubes o 8.7 mm diameter, 1.88 m length per pass and 90 m circuit total length. On the air side, corrugated aluminium ins o rectangular shape were set at 197 ins per meter. Air enters at the evaporator ront at -4 C; rerigerant CO enters at a saturation temperature o -30 C and a quality o X=0 rom the rear o the coil. When CO reaches the end o the circuit (at the coil ront) the quality becomes X=1 (complete evaporation). Average air temperature distributions or each tube pass are represented in Fig.3. Temperature along each tube row is uniorm and at the tube exit, temperature is constant and equal to -7.4 C. It is interesting to note in this case that besides this uniormity o temperature behind each tube row, the air temperature drop across tube rows, starting with a relatively small average value o 0.5 o C or the irst two rows, decreases gradually down to approximately 0. o C towards the last rows. Due to the thermodynamic properties o carbon dioxide, particularly its high saturation pressures or the temperatures levels used in rerigeration, the temperature decrease accompanying pressure losses inside the tube is
modest. Overall, this act coupled with the controlling eect o heat transer by the air side due to its high thermal resistance, results in a balanced air temperature distribution across the coil. This gradual decrease with a small temperature gradient avours a good rerigeration load repartition over the Fig. 3 : Distribution de la température moyenne de l air Fig.: Experimental set-up schematics entire heat transer surace, or given rerigeration conditions. In case o rost deposition it is expected to be more uniorm and to occur over a longer period o time than or an ordinary synthetic rerigerant. The distribution o the cumulative pressure losses as a unction o the tube length is presented in Fig.4. It is worth noting that 80% o the total losses occur in the irst hal o the circuit, corresponding to increasing qualities, vapour velocities and better heat transer. Over the total length o 90 meters, including 80 return bends (180 o bends), the total pressure drop losses or CO amounts only to 80 ka. This act highlights one the many advantages resulting rom the avourable thermo physical and transport properties o CO. This pressure drop results in the diminution o the saturation temperature by 1.8 C as can be seen in Fig. 5. Fig.4: Cumulative pressure losses o CO For the sake o comparison with the synthetic rerigerants ordinarily used in rerigeration systems, this temperature reduction value is rapidly attained in relatively short tube lengths. In order to maintain a reasonably constant temperature across an evaporator this temperature drop must be small (less than 1 o C say). This condition needs several small length circuits with synthetic rerigerants, while ar ewer circuits may be required with carbon dioxide. The distribution o the air temperature shown in Fig.5 and represented by blank squares has eight lat plateaus, corresponding to the number o rows deep. Since these plateaus are essentially horizontal, the assumption o an average temperature beore each pass is justiied. In order to outline the advantages oered by the thermo physical properties o CO, a comparative study with 5
Fig.5: Temperature distribution o CO, air and wall Fig. 7: Temperature distribution o air and R507A CONCLUSION Fig. 6: Four circuits coniguration or the study o R507A Rerigerant R507 has been carried out. Several iterative tests have been perormed in order to obtain a reasonable temperature drop. It was ound out to be impossible to use less than our circuits. Thereore the coniguration selected was that o our circuits as shown in Fig.6. In such a case the circuits are well balanced and the temperature drop in the saturation temperature is o the order o.4 o C in each circuit. Air temperature distribution also presents eight plateaus. In contrast with CO however, these plateaus are not horizontal as Fig. 7 shows, a act that explains the lack o uniormity in the air temperature beore each pass. A theoretical model allowing simulation o an air/co evaporator coil with corrugated aluminium ins has been developed and validated. The structure o the resulting tool is such that details on the temperature distribution, vapour quality, CO pressure and other parameters along the circuit can be obtained. The results rom these computations allow or detailed inormation on the luids states in the coil. On the air side, temperature, relative humidity and external pressure drop are among the important parameters that can be obtained but only representative results have been presented in this paper or the sake o conciseness. On the rerigerant side, saturation temperature, internal pressure drop and vapor quality along the coil can be tracked. Local results can be integrated to obtain global values or engineering purposes. It is possible to qualiy tube rows by determining their individual capacities as a unction o their location in the coil. Experiments have been perormed on an experimental acility o the CTEC-V Research Centre. Results rom this installation as well as those collected rom the open literature have been used to validate the developed model. Agreement between the model predictions and the experimental values is satisactory. As an example, calculations have been perormed on an evaporator coil typically used in rerigerated cabinets with three doors. The coil capacity is approximately 3.5 kw and the air inlet temperature is -4 C. Comparison with rerigerant R507A shows that in general several circuits are necessary, while or CO a single circuit coniguration was suicient to obtain the same or even smaller temperature drop, which was acceptable in rerigeration practice. More speciically, over a tube length o 90 meters, the temperature drop was ΔT ev = 1.8 C. By means o this model some advantages oered by the 6
thermo physical properties o this rerigerant have been demonstrated, among others: - Air temperature ater every tube pass is nearly constant. - ressure drop or CO is very small in comparison to other rerigerants ordinarily used in current systems and results in correspondingly small drop in saturation temperature. This has a positive impact on the temperature distribution within the reezing cabinet. - It is possible to use longer circuits with CO (in comparison to other rerigerants), thereore reducing their number or a given capacity. This greatly simpliies the coil geometry, impacting positively on rerigerant distribution and the cost o the coils. ACKNOWLEDGEMENTS This work has been supported by the rogram o Energy Research and Development (ERD) o Natural Resources Canada. NOMENCLATURE A area, m Cd centre to center distance o bend Cp speciic heat, J/kg o C D diameter, m Fs in spacing, m riction actor G mass lux, kg/m o C GF geometry actor h convective heat transer coeicient, kw/m o C H enthalpy, kj/kg J colburn heat transer actor L length, m m mass low rate, kg/s N number o ins N ligne height coordinate N row row number l longitudinal tube pitch t transversal tube pitch pressure, a Q capacity, kw Re Reynolds number T temperature, o C U g overall heat transer coeicient, W/m o C v speciic volume, m 3 /kg x quality ΔL element o length, m Δ pressure drop, a ΔT temperature drop, o C λ thermal conductivity o tube, kw/m o C η eiciency ω humidity ratio, kg vapor/kg total moist air ρ density, kg/m 3 θ corrugated angle, degrees Subscripts a air b bare bc convective boiling ben bend c collar co CO Dc collar diameter ev evaporation ins g gas h hydraulic in inner l liquid m mean nb nucleate boiling ou outer o overall s singular sat saturation T tube t transversal tp two phase v vapor w wall wi internal wall REFERENCES Geary, D.F., 1975, Return Bend ressure Drop in Rerigeration Systems, ASHRAE Transaction, Vol. 81, (1), pp. 50-64. Inlow, S.W. and Groll, E. A., 1996, Analysis o Secondary Loop Rerigeration Systems Using Carbon Dioxide as Volatile Secondary Rerigerant, HVAC&R Research, Vol., no., pp. 107-10. Jiang, H., Aute, V. and Radermacher, R., 006, Coil Designer: A General-urpose Simulation and Design Tool or Air-to-Rerigerant Heat Exchangers, Int. J. o Rerigeration, Vol. 9, pp. 601-610. Liang, S.Y. Liu, M., Wong, T. N. and Nathan, G. K., 1999, Analytical Study o Evaporator Coil in Humid Environment, Applied Thermal Engineering, Vol.19, pp.119-1145. Lorentzen, G., 1995, The use o natural rerigerants: a complete solution to the CFC/HCFC predicament, Intl. J. o Rerigeration, Vol.18 (4), pp.190-197. Ouzzane, M. and Aidoun, Z., 008, A numerical Study o a Wavy Fin and Tube CO Evaporator Coil, to appear in Heat Transer Engineering. Wang, C. C., Hwang, Y. M. and Lin, Y. T, 00, Empirical Correlations or Heat transer and Flow Friction Characteristics o Herringbone Wavy Fin and Tube heat Exchangers, Int. Journal o rerigeration, Vol. 5, 673-680. ASHRAE, 1993, ASHRAE Handbook o Fundamentals, SI Edition, Atlanta. Rohsenow, W. M, Hartnett, J.. and Cho, Y. I., 1998, Handbook o Heat Transer, Third Edition, Mc Graw Hill, New York. Aidoun, Z., Ouzzane, M., 004, A Study o Evaporator Coil Conigurations Using CO in a Rerigeration Loop,Natural Working Fluids, roc. 6 th IIR- Gustav Lorentzen Conerence, 7
August 9-September1 st, 004, Glasgow, United Kingdom, 4/A/1.40. Dispenza, A., Dispenza, C., La Rocca, V., anno, D. and anno G., 005, Advanced Rerigerating lants Based on Transcritical Cycles Working With carbon Dioxides or Commercial Rerigeration, roc. IIR International Conerences, Commercial Rerigeration, 31 Aug- Sept. 005, Vicenza, Italy, pp. 195-0. Hwang, Y. B., Kim, H. and Radermacher, R., 1997, Boiling Heat Transer Correlation or Carbon Dioxide, roc. IIR International Conerence, Heat Transer Issues in Natural Rerigeration, Nov. 1997, Maryland, pp. 81-95. Sawalha, Rogstam, Nilsson, S. J. and. O., 005, Laboratory tests o NH 3 /CO Cascade System or Supermarket Rerigeration, roc. IIR International Conerences, Commercial Rerigeration, 31 Aug- Sept. 005, Vicenza, Italy, pp. 11-18. earson, S.F., 1993, Development o Improved Secondary Rerigerants, roceedings, Institute o Rerigeration, March 1993, U.K., paper 7, pp. 65-80. Lorentzen, G., 1993, Application o natural rerigerants, IIR Annex 1993, Gent, pp. 55-64. McLinden, M.O., Klein, S.A., Lemmon, E.W., eskin, A.., 1998, NIST Thermodynamic and Transport roperties o Rerigerant Mixtures-REFRO, Version 6.01, U.S. Department o Commerce, Technology Administration, Gaithersburg, Maryland, July. 8