Indirect Evaporative cooling systems: modelling and performance analysis

Available online at www.sciencedirect.com ScienceDirect Energy Procedia 140 (2017) 475 485 www.elsevier.com/locate/procedia AiCARR 50t International Congress; Beyond NZEB Buildings, 10-11 May 2017, Matera, Italy Indirect Evaporative cooling systems: modelling and performance analysis Paolo Liberati a, *, Stefano De Antonellis b, Calogero Leone a, Cesare Maria Joppolo b, Yakub Bawa a a Recuperator S.p.A., Rescaldina (MI), Italy b Politecnico di Milano, Dipartimento di Energia, Milano (MI), Italy Abstract Nowadays, tere are a lot of ongoing researc activities about cooling tecnologies wic can lead to a significant reduction in primary energy consumption of air andling units, in bot residential and commercial applications. In tis contt, one of te most interesting tecnologies is te indirect evaporative cooling (IEC) system. In tis work a penomenological model of te component, based on a cross flow eat canger, as been developed and validated in typical summer operating conditions, at different air streams temperature, umidity ratio and flow rates. Using tis tool, performance of te IEC system as been analyzed in different working conditions. Results igligt te advantage of using te IEC unit, wic leads to significant energy savings in almost all te investigated conditions. 2017 Te Autors. Publised by Elsevier Ltd. Peer-review under responsibility of te scientific committee of te AiCARR 50t International Congress; Beyond NZEB Buildings. Keywords: indirect evaporative cooling, performance, cross flow, eat canger; 1. Introduction Te indirect evaporative cooling tecnology (IEC) is a solution tat can be used in te air andling units to reduce te energy consumption during te summer period, bot in residential and commercial applications. * Corresponding autor. Tel.: +39 0331 1853 1. E-mail address: liberati@recuperator.eu 1876-6102 2017 Te Autors. Publised by Elsevier Ltd. Peer-review under responsibility of te scientific committee of te AiCARR 50t International Congress; Beyond NZEB Buildings 10.1016/j.egypro.2017.11.159

476 Paolo Liberati et al. / Energy Procedia 140 (2017) 475 485 In tis system, te aust air, at ambient conditions, is umidified wit water sprayed by te nozzles installed in proximity of te eat canger entrance; at te same time, te supply air, at outdoor temperature and umidity, is cooled down by te fres and umid air of te opposite side. Nowadays tere are a lot of ongoing researc activities about te IEC solutions; at te moment, main works ave focused on studying prototypes [1, 7, 8] or particular operative conditions [5, 6]. Anyway tere aren t works analyzing performance of IEC systems based on cross flow plate eat cangers in typical summer conditions. Nomenclature A HE,net Heat canger net cross section area, m 2 cp Specific eat, J kg -1 K -1 Cw Wettability coefficient, - f eva Fraction of te adsorbed water, - Net cannel eigt, m M Convective mass transfer coefficient, kg s -1 m -2 T Convective eat transfer coefficient, W K -1 m -2 k W Plate conductivity, W m -1 K -1 L Net eat canger lengt, m m Specific mass flow rate, kg s -1 m -2 N HE Number of eat canger plates, - pt Heat canger plates pitc, m Flow rate, m³ -1 w Water flow rate sprayed on te eat canger, l -1 T Temperature, C v Velocity, m s -1 x Primary air flow direction, m X Humidity ratio, kg kg -1 y Secondary air flow direction, m Greek symbols δ Heat canger plates tickness, mm ΔT Temperature difference, C ε Heat canger effectiveness, - ε wb Wet bulb IEC effectiveness, - ρ Density, kg m -3 σ Wettability factor, - φ Relative umidity, - Subscript a Air Exaust air HE Heat Excanger in Inlet out Outlet su Supply air W Wall eat canger plates wb Wet bulb temperature Superscript N In reference conditions (ρ = 1,2 kg m -3 )

Paolo Liberati et al. / Energy Procedia 140 (2017) 475 485 477 Tis researc aims to fulfill tis lack studying te IEC system in different operating conditions (namely air streams temperature, umidity ratio and flow rates). Te work is divided into tree parts. Te first part deals wit te description of te perimental setup and te test rig adopted in te researc. Te second part plains te developed model and its calibration process. Te last part igligts te advantage of using te IEC unit troug a parametric analysis based on te developed model. It is sown te indirect evaporative cooling tecnology leads to significant energy savings in all te investigated conditions, even wit a low water flow rate. 2. IEC SYSTEM DESCRIPTION A sceme of te analyzed indirect evaporative cooler is sown in Figure 1. Te system consists of: A commercial cross-flow plate eat canger. N 8 water spray nozzles. Te equipment to increase pressure of water supplied to te nozzles. Te supply air stream is cooled in te eat canger at constant umidity ratio: it enters te system in condition su,in (assumed at outdoor air conditions) and it leaves te component in condition su,out. Te aust air stream enters te system in condition,in (assumed at indoor air conditions - return air stream from te building) from te top plenum were water nozzles are installed. Due to te evaporation of water, te air stream is umidified almost at constant entalpy from condition,in (before te plenum inlet) to condition,in,he (at eat canger face). Finally, te secondary air stream leaves te system at condition,out. According to Figure 1 and to te actual perimental setup, te supply air stream crosses te eat canger from te rigt to te left and te aust air flow from te top to te bottom of te system. Main data of te eat canger adopted in tis study are summarized in Table I. Table 1.Main data of te investigated IEC system Description Parameter Value Number of plates N HE 119 Plate tickness δ 0.14 mm Plate pitc pt 3.35 mm Net cannel eigt =pt- δ 3.21 mm Net plate lengt and widt L* 470 mm Plate conductivity k W 220 W m -1 K -1 Water nozzles are installed on two parallel manifolds (n 4 nozzles on eac one). Te distance between eac nozzle is around 8 cm and te two manifolds are installed at 15 cm from te eat canger face. According to data provided by manufacturer, te nominal water flow rate of eac axial flow - full cone nozzle is 7.50 l -1 at 9 bar. Nozzles are installed in order to provide water in counter current arrangement respect to te secondary air stream. Te lengt of top and side plenums is 42 cm and te lengt of bottom plenum is around 90 cm. Inlet temperature, umidity ratio and flow rate of eac stream are set troug a dedicated air andling unit. Temperature and relative umidity are measured at te inlet and outlet of te investigated IEC system troug RTD PT100 sensors (± 0.2 C at 20 C) coupled to capacitive relative umidity sensors (± 1% between 0 and 90%). A detailed description of te perimental setup is available in previous works of te autors [2, 3]. 3. IEC SYSTEM MODELING AND VALIDATION Te model of te indirect evaporative cooling system as been described and discussed in detail in a previous work of te autors [4]. Terefore, in tis paper main assumptions and equations are briefly reported.

478 Paolo Liberati et al. / Energy Procedia 140 (2017) 475 485 Main adopted assumptions are: Fig. 1. Sceme of te perimental setup Steady-state conditions. No eat losses to te surroundings. Negligible axial eat conduction and water diffusion in te air streams. Negligible eat conduction in te eat canger plates. Uniform air inlet conditions. Interface plate temperature equal to bulk water temperature. Referring to Fig. 2, energy and water mass balances ave been applied to an infinitesimal element of te eat canger: dt UT, su TW T su su dx vsucpsusu 2 T T cp T X X dt T, W dy vcp 2 dx dy M, X W X v 2 M, W vcp 2 cp T X X T T U T T 0 (4) M, W T, W T, su su W M, X X W (5) dmw dy / 2 Were σ is te fraction of wet surface area and U T,su = 1/(1/ T,su +δ/kw). Te mass transfer coefficient is calculated assuming Le = 1 as: (1) (2) (3)

Paolo Liberati et al. / Energy Procedia 140 (2017) 475 485 479 M T cp a (6) According to te results reported in a previous work of te autors for te same eat canger [4], te eat transfer coefficient T is calculated troug te following correlation: k a T 2 0.0185 Re 0.928 Pr 1/ 3 (7) Were k a is te termal conductivity of air and 2 is te ydraulic diameter of te cannel. According to [4], te correlations to predict te saturation efficiency, te wettability factor and te coefficient C w are respectively: c T, in T, wb, in c2 4 M 1 ln c c w, in 3 M s (8) m v m C w w w 2 w w w (9) C w v N k 2 S k e 1 k3 m w, in, HE Were m w,in,he = m w,in - (X,in,HE - X,in ) N ρn /(3600 A HE,net ) is te water specific mass flow rate, net of te water evaporation in te secondary air inlet plenum, and m w,in = M w,in /A HE,net, wit te net eat canger cross area A HE,net equal to 0.089 m 2. Due to te different plenum geometry adopted in tis researc, compared to te previous study of te autors [4], te parameters of Eqs. (8 10) ave been fitted wit perimental data (n 30 tests), as summarized in Table 2. As a result it is: c 1 = -6.781, c 2 = 33.97, c 3 = 0.954, c 4 = 0.984, k 1 = 4.821, k 2 = 0.0114 and k 3 = 4.819. In Fig. 3, te parity plot comparing perimental and numerical wet bulb effectiveness is reported for te data adopted in te calibration process (Table 2): in all cases te difference is witin 3%. Te effectiveness is defined as: (10) ε wb = (T su,in -T su,out ) / (T su,in -T wb,,in ) (11) Fig. 2. Sceme of control volume adopted in te model

480 Paolo Liberati et al. / Energy Procedia 140 (2017) 475 485 Fig. 3. Parity plot of wet bulb effectiveness (Data of Table 2) Table 2. Main perimental data used for IEC model calibration Test T,in X,in T su,in X su,in w,in [ C] [g kg -1 ] [m 3-1 ] [ C] [g kg -1 ] [m 3-1 ] [l -1 ] A 30.0 10.6 1200,1800 35.0 10.0 1200 30-60 B 30.0 13.4 1200,1800 35.0 10.0 1200 30-60 C 36.8 10.6 1200,1800 35.0 10.0 1200 30-60 4. IEC performance: results discussion N In Fig. 4 to 6 te effect of different airflows, aust and supply air conditions and water flow rate on system performance is evaluated. Te wole following analysis ave been carried out considering te same plate eat canger and nozzles configuration described in section 2 and 3. Tree conditions of water flow rate ave been analyzed: no water flow (dry condition), w,in = 15 l/. w,in = 30 l/. Tree combinations of supply-aust airflow rate: su N = 1200 m³/ and N = 1200 m³/. su N = 1800 m³/ and N = 1800 m³/. su N = 1800 m³/ and N = 1200 m³/. Comparison between perimental data as been performed troug te following indes: primary air temperature difference and fraction of evaporated water. Suc quantities are defined in tis form: su N ΔT su = T su,in - T su,out (12) f eva = N ρ a N (X,out - X,in ) / (M w,in 3600) (13)

Paolo Liberati et al. / Energy Procedia 140 (2017) 475 485 481 Results of te simulation wit balanced airflow at 1200 m³/ are sown in Fig. 4: te water sprayed on te plate eat canger as a strong influence on te cooling capacity of te system; te iger is w,in, te iger is ΔT su in all te analyzed conditions. Compared to te dry condition (at T su,in = 34 C), in case of water flow rate of 15 l/ te cooling capacity almost doubles; wen w,in = 30 l/ te cooling capacity increases around 120%. Te evaporation of te water layer on eat canger plates, cools down te temperature of te plate surface and, consequently, increases significantly te system cooling capacity. Fig. 4. Temperature difference between te inlet and outlet of te supply air flow in function of te outdoor temperature wit w,in =0 l/ (A), w,in =15 l/ (B), w,in =30 l/ (C) and fraction of te evaporated water wit w,in =15 l/ (D), w,in =30 l/ (E) for 4 different aust T and φ. su N =1200 m³/ N =1200 m³/

482 Paolo Liberati et al. / Energy Procedia 140 (2017) 475 485 Fig. 5. Temperature difference between te inlet and outlet of te supply air flow in function of te outdoor temperature wit w,in =0 l/ (A), w,in =15 l/ (B), w,in =30 l/ (C) and fraction of te evaporated water wit w,in =15 l/ (D), w,in =30 l/ (E) for 4 different aust T and φ. su N =1800 m³/ N =1800 m³/ It sould be noticed tat te effect on te performance is greater in te range from w,in = 0 to w,in = 15 l/ tan in te range from w,in = 15 l/ to w,in = 30 l/. Tis is because te effect of te indirect evaporative cooling as an asymptotic beavior: at ig water flow rates a ig fraction of te eat canger surface is wet and te air stream reaces almost saturation conditions. In te Fig. 4B different aust air conditions are compared. At constant outdoor temperature, ΔT su rises wit te decreasing of relative umidity and temperature of te aust airflow: a lower dry bulb temperature of te aust air increases te sensible cooling capacity and a lower relative umidity promotes water evaporation. During te IEC process only part of water droplets evaporates in te aust air stream. Te fraction of evaporated water

Paolo Liberati et al. / Energy Procedia 140 (2017) 475 485 483 increases wit te increase in T su,in and te term T,in -T wb,,in. Quite obviously te ΔT su is iger for te condition T,in = 24 C φ,in = 40% tan T,in = 26 C φ,in = 40% despite te fraction of te evaporated water is similar in bot conditions. In te condition T,in = 26 C φ,in = 40% te fraction of te evaporated water varies from f eva = 0.4 to f eva = 0.6. Tis value decreases wit te increase of te water flow rate: wen w,in = 30 l/ te fraction of te evaporated water is always lower tan f eva = 0.4, even wit ig outdoor temperatures (T su,in ). In Fig. 5 it can be noticed te effect of te increase in te velocity of bot supply and aust airflows. Te increase in te airflow from Q N = 1200 m³/ to Q N = 1800 m³/ does not lead to a significant cange in cooling performance bot in dry condition (Fig. 5A) and in wet condition (Fig. 5B). Fig. 6. Temperature difference between te inlet and outlet of te supply air flow in function of te outdoor temperature wit w,in =0 l/ (A), w,in =15 l/ (B), w,in =30 l/ (C) and fraction of te evaporated water wit w,in=15 l/ (D), w,in =30 l/ (E) for 4 different aust T and φ. su N =1800 m³/ N =1200 m³/

484 Paolo Liberati et al. / Energy Procedia 140 (2017) 475 485 Comparing to Fig. 4, in dry condition only a sligt decrease ΔT su occurs. In te wet condition, te increase in te airflow rate is partially compensated by te increase of te quantity of evaporated water; for instance, te fraction of te evaporated water reaces te value of 0.65 wen T,in = 26 C, φ,in = 40% and T su,i n = 34 C (Fig. 5D), wic is 35% iger tan in te condition reported in Fig. 4D. In te AHU system te supply airflow rate is often iger tan te aust air flow rate. Te performance of te IEC system in condition wit unbalanced flows is sown in Fig 6: te supply air is equal to su N =1800 m³/ wile te aust air is N = 1200 m³/. In tis condition ΔT su decreases around two degrees compared to results of Fig 4. On te contrary, te quantity of evaporated water, sown in Fig. 6D, is iger tan te one of Fig. 4D because te average temperature of te plate is iger due to te iger supply air flow. Te quantity of supplied water on te system influences strongly te cooling capacity of te IEC (Fig. 7). At constant condition, wit low water flow rate, te ΔT su rises steeply but it tends to an asymptotic value at ig water flow rate. At a ig water flow rate ( w,in > 35 l/), te cooling capacity sligtly increases wile te fraction of te evaporated water decreases consistently. Fig. 7. Temperature difference between te inlet and outlet of te supply air flow in function of te water sprayed (A) and fraction of te evaporated water (B) for 4 different aust T and φ. su N = 1200 m³/ N =1200 m³/. T su,in=34 C 5. Conclusion In tis work, performance of an indirect evaporative cooling system as been discussed. A penomenological IEC system model as been developed and calibrated wit perimental data collected in a dedicated test facility. Wit te model, te IEC system as been analyzed and significant results arose. Te water sprayed on te eat canger strongly increases te system cooling capacity even wit very low flow rate: in te investigated conditions, wen w,in=15 l/ te cooling capacity can be twice te one in dry conditions.

Paolo Liberati et al. / Energy Procedia 140 (2017) 475 485 485 Te fraction of te adsorbed water depends on te temperature and umidity condition of te two airflows, tis value can rise up to f eva =0.7 wit ig supply and aust flow rate and low water flow rates. Terefore, IEC systems can be an effective tecnology to acieve significant primary energy savings in HVAC operating in summer conditions. References [1] Bruno F. 2011. On-site perimental testing of a novel dew point evaporative cooler. Energy and Buildings, 43 (12), 3475-3483 [2] De Antonellis S., Intini M., Joppolo C.M., Molinaroli L., Romano F. 2015. Desiccant weels for air umidification: An perimental and numerical analysis. Energy Conversion and Management, 106, 355-364 [3] De Antonellis S., Joppolo C.M., Liberati P., Milani S., Molinaroli L. 2016. Experimental analysis of a cross flow indirect evaporative cooling system. Energy and Buildings, 121, 130-138 [4] De Antonellis S., Joppolo C.M., Liberati P., Milani S., Romano F. 2017. Modeling and perimental study of an indirect evaporative cooler. Energy and Buildings, 142, 147-157 [5] Heidarinejad G., Bozorgmer M., Delfani S., Esmaeelian J. 2009. Experimental investigation of two-stage indirect/direct evaporative cooling system in various climatic conditions. Building and Environment 44 (10) 2073-2079 [6] Kim M.-H., Kim J.-H., Coi A.-S., Jeong J.-W. 2011. Experimental study on te eat cange effectiveness of a dry coil indirect evaporation cooler under various operating conditions. Energy, 36 (11), 6479-6489 [7] Saman W.Y., Alizade S. 2002. An perimental study of a cross-flow type plate eat canger for deumidification/cooling. Solar Energy, 73 (1), 59-71 [8] Tejero-González A., Andrés-Cicote M., Velasco-Gómez E., Rey-Martínez F.J. 2013. Influence of constructive parameters on te performance of two indirect evaporative cooler prototypes. Applied Termal Engineering, 51 (1-2), 1017-1025 [9] Zang H., Sao S., Xu H., Zou H., Tian C. 2014. Free cooling of data centers: A review. Renewable and Sustainable Energy Reviews, 35, 171-182 [10] Ziyin D., Cangong Z., Xingxing Z., Mamud M., Xudong Z., Berang A., Ala H. 2012, Indirect evaporative cooling: Past, present and future potentials, Renewable and Sustainable Energy Reviews 16 (9) 6823-6850