A Novel Evaporative Cooling System with a Polymer Hollow Fibre Spindle

: A novel evaporative cooling system, in which the hollow fibre module constitutes as the 14 humidifier and evaporative cooler, is proposed. With the aim to avoid the flow channelling or shielding 15 of adjacent fibres the fibres inside each bundle were made into a spindle shape to allow maximum 16 contact between the air stream and the fibres. This novel hollow fibre integrated evaporative cooling 17 system will provide a comfortable indoor environment for hot and dry area. Moreover, the water vapour 18 can permeate through the hollow fibre effectively, and the liquid water droplets will be prevented from 19 mixing with the processed air. Under various inlet air dry bulb temperatures (27◦C, 30◦C, 33◦C, 36◦C 20 and 39◦C), and various inlet air relative humidity (23%, 32% and 40%), the cooling performances of 21 the proposed novel evaporative cooling system were experimentally investigated. The variations of 22 outlet air dry bulb temperature, wet bulb effectiveness, dew point effectiveness and cooling capacity 23 with respect to different incoming air dry bulb temperature were studied. The effects of various 24 incoming air Reynolds number on the heat and mass transfer coefficients, heat flux and mass flux across 25 the polymer hollow fibre module were analysed. Experimentally derived non-dimensional heat and mass transfer correlations were compared with other correlations from literature. Due to the spindle shape of proposed hollow fibre module, the shielding with hollow fibre bundles could be avoided greatly, therefore the mass transfer performance of the proposed system demonstrated significant improvement compared with other devices reported in literature.

the polymer hollow fibre module were analysed. Experimentally derived non-dimensional heat and 26 mass transfer correlations were compared with other correlations from literature. Due to the spindle 27 shape of proposed hollow fibre module, the shielding with hollow fibre bundles could be avoided 28 greatly, therefore the mass transfer performance of the proposed system demonstrated significant 29 improvement compared with other devices reported in literature. Heat flux (W/ m 2 ) 46 0.35m 2 -1.13m 2 , and around 0.4•C temperature drop could be observed from the experiments. The above 119 publications are mainly concentrated on the theoretical analysis on the polymer hollow fibre integrated 120 evaporative cooling system. The available experimental results were limited to the variation of outlet 121 air temperatures with respect to different air flow rates. In addition, as stated by Johnson et al.[15], due 122 to the shielding from adjacent fibres, the heat and mass transfer performance will decrease when using 123 a large number of fibres inside one module. 124 A summary of the recent experimental and modelling works on evaporative cooling system is presented 125 in Table 1. Literature review indicates that the previous published papers were mainly concentrated on 126 the theoretically modelling of evaporative cooling system. For the limited experimental investigations 127 reported in the literature, the evaporative coolers were mainly made from porous paper materials. This 128 paper presents a novel evaporative cooling system with a hollow fibre evaporative cooler. Instead of 129 previously reported cross flow configurations [13,15,21], five fibre bundles (each contains 100 fibres) 130 with the distance of 5cm were placed normal to the air stream, with detailed configuration shown in 131 Figure 2. In order to avoid the flow channelling or shielding of adjacent fibres, the fibres inside each 132 bundle were made into a spindle shape to allow maximum contact between the air stream and the fibre. 133 As a subsequent work of previous research [22], this research work extends the previous experimental 134 testing conditions to a wider range, with the incoming air temperature up to 39•C and relative humidity 135 up to 40%.
The variations of outlet air dry bulb temperature, wet bulb effectiveness, dew point 136 effectiveness and cooling capacity were studied by varying the incoming air dry bulb temperature from 137 27•C to 39•C and RH from 23% to 40%. The effects of various incoming air Reynolds number on the 138 heat and mass transfer coefficients, heat flux and mass flux across the polymer hollow fibre module 139 were analysed. Two sets of experimentally derived non-dimensional heat and mass transfer correlations 140 were summarized, which could be favourable for the future design of polymer hollow fibre integrated 141 evaporative cooling system. 142

Heat and mass transfer of hollow fibre evaporative cooling system 143
The 3D model of the hollow fiber module is shown in Figure 1 together with the temperature 144 and humidity change profile. As illustrated in Figure 1, the incoming hot and humid air gets 145 in contact with the porous hollow fiber module, inside which the water will be circulating 146 around. As water evaporates through the hollow fibers, it will extract energy from the 147 incoming air causing the temperature to drop. As reported by Johnson et al.[15], the 148 resistance of fibe materials is very small, therefore it could be neglected. According to 149 Rawangkul et al[5], 150 The sensible cooling capacity (Q ) supplied by the incoming air of such novel evaporative cooling 151 system can be calculated by: 152

Eq.(1) 153
Where, Q is the flow of transferred heat (W); 154 m ୟ is the mass flow rate of the incoming air (kg/h);

155
‫ܥ‬ is the specific heat of air at constant pressure, kJ/(kg K); 156 T ଵ is the dry bulb temperature of the incoming air (•C); 157 T ଶ is the dry bulb temperature of the outgoing air (•C);

158
ߩ is the density of the air, kg/m 3 ; 159 V is the volumetric flow rate of the incoming air, m 3 /h; 160

161
The rate of watere evaporation is: Where ݉ is the flow of evaporated water(kg/h); m ୟ is the mass flow rate of the incoming air (kg/h); 165 ߱ ଵ is the humidity ratio of the incoming air; 166 ߱ ଶ is the humidity ratio of the outgoing air.

167
The rate of heat transfer (Q) and the rate of water evaporation (m ୣ ) can be given as the 168 product of heat transfer coefficient and the mean logarithmic difference in temperature (ΔT), 169 and the product of the mass transfer coefficient and the mean logarithmic difference in the 170 water vapour density (Δρ ). These can be expressed in the following two equations: Where h ୌ is the coefficient of heat transfer (W/m 2 K);

174
h is the coefficient of mass transfer (W/m 2 K); 175 q is the heat flux (W/m 2 ); 176 N is the mass flux (mg/ m 2 s);

177
A ୱ is the total surface area of the polymer hollow fibre (m 2 );

178
The mean logarithmic difference in temperature (ΔT) and water vapour density (Δρ ) can be 179 calculated using following equations: Where ρ ଵ and ρ ଶ are the water vapour density on entering and leaving the hollow fibres 183 (kg/m 3 ); 184 ρ ௪ is the saturated water vapour density at the wet bulb temperature (kg/m 3 ).

185
The heat and mass transfer coefficients could be calculated from: Where k is the thermal conductivity of the air (W/mK);

189
݀ is the hydraulic diameter of the hollow fibre module, m. According to [23], ݀ can be 190 calculated as : 191 According to the hollow fiber bundle configuration [24], ݀ can further be expressed as: Where ‫ܥ‬ ଵ , ‫ܥ‬ ଶ , ݉ ଵ , ݉ ଶ are constants for the hollow fibre bundles.

203
Reynolds number can be calculated by: Where ‫ݑ‬ is the incoming air velocity, m/s; 206 ν is the viscosity of the incoming air, m 2 /s.

207
From the previous study [22], using the experimental data, general empirical correlations for the non-208 dimensional heat and mass transfer data for the proposed system are derived using mathematical data 209 regression techniques. In this research, as the experimental testing conditions are extended to a large 210 range, the applicability of these equations will be verified in section 4. The wet bulb effectiveness (ε ௪ ) is an important expression used to characterise the air saturation 214 capacity of the polymer hollow fibre bundle. This is defined as the ratio between the thermal 215 difference on passing through the hollow fibre bundle (ܶ ଵ − ܶ ଶ ) and the maximum thermal difference 216 that would occur if the air were saturated (ܶ ଵ − ܶ ௪ ): 217 Similarly, the dew point effectiveness (ε ௗ௪ ) is defined as the ratio between the incoming air and the 219 outgoing air to the difference between the incoming air and its dew point temperature, as indicated by 220 the following expression: 221

Eq.(17) 222
A lab scale experimental testing rig is developed, which integrates the hollow fiber evaporative cooler 224 with the evaporative cooling system. Polyvinylidene fluoride (PVDF) hollow fibres (manufactured by 225 ZENA Ltd.) with outside diameter of 0.8mm and inside diameter of 0.6mm, an effective pore size of 226 0.5µm and a porosity of 50% were used for the fabrication of the polymer hollow fibre module. The 227 polymer hollow fibre module consists of 5 fibre bundles (each contains 100 fibres), which were 228 connected at each bundle ends using T piece plastic tubing. The hollow fibre module was incorporated 229 into a circular aluminium tunnel, whose cross section diameter was 0.15m. In order to avoid the flow 230 channelling or shielding of adjacent fibres, the fibres in each bundle were compressed from both ends 231 to make the bundle into a spindle shape to allow maximum contact between the air stream and the fibres. 232 The detailed physical properties of the polymer hollow fibre module were summarized in Table 2. 233 The testing rig consists of a polymer hollow fibre module, an air tunnel, a water pump, a fan and a water 234 tank. The detailed schematic diagram is shown in Figure 2. A 5-litre water tank was used to provide 235 water circulation inside the fibre. In order to avoid any particle blockage within the polymer hollow 236 fibres, a water filter was allocated to improve the purity of the incoming water into the fibres. A flow 237 meter and a ball valve were included in the water circulation cycle with the aim to control the water 238 flow rate inside the fibre. In order to minimize the experimental testing errors, four humidity and 239 temperature sensors (EK-H4, Sensirion, UK) were located at the inlet (point 1 in Figure 2) and outlet 240 (point 2 in Figure 2) of the tunnel respectively, to measure the inlet and outlet conditions of the air 241 stream. Additional K type thermocouples were used to measure the water temperature entering and 242 leaving the hollow fibre module (point 3 and 4 in Figure 2). The aluminium tunnel was connected with 243 a variable frequency drive centrifugal fan, which was linked directly with the environmental chamber. 244 The experimental prototype image is shown in Figure 3. The basic working principle of the experiment 245 is as following: The hot and humid air from the environmental chamber will be blown out by the blower 246 into the air tunnel. With the help of the circulation pump, water is circulating from the water tank into 247 the hollow fibre bundles to ensure that the fiber surface will get wetted throughout the tests. The 248 incoming hot and humid air will get in contact with the hollow fibre module. As water evaporates 249 through the porous hollow fibres, the temperature of incoming air will be reduced, accompanied by the 250 increased relative humidity. 251 At the beginning of each test, the environmental chamber was set to the required temperature and 252 humidity level. As soon as the temperature and humidity reached the desired values, the water pump 253 for the fibre module and the fan in the air stream direction was switched on. The air velocity was 254 measured at the five different positions along the cross sections of the outlet aluminium tunnel, using 255 the air velocity probes connected to a recorder (Testo 454). The dynamic pressures in the upstream and 256 downstream of the air flow were recorded using Pressure transducers (Ge UNIK 5000). 257 For each test, the temperature and humidity values were recorded every 20 seconds until the time when 258 the system reach steady states as indicated by the humidity and temperature sensor readings. The 259 accuracy of the measuring instruments used was: ±0.2% for temperature, ±0.5% for pressure, ±2% for 260 air velocity, and ±2% for relative humidity. 261

Results and discussion 262
As indicated in Table 3, the experiments were carried out for the proposed novel evaporative cooling 263 system with various inlet air dry bulb temperatures (27•C, 30•C, 33•C, 36•C and 39•C), and various 264 inlet air relative humidity (23%, 32% and 40%). Figure 4 presents the variation of outlet air dry bulb 265 temperatures with respect to various inlet air dry bulb temperatures under different incoming air 266 relative humidity. For incoming air temperature in the range of 27-39•C and the incoming air relative 267 humidity varying from 23%, 32% to 40%, it can be observed that the outlet air temperature is 268 dramatically affected by inlet air relative humidity at constant incoming air temperature. At the same 269 inlet air dry bulb temperature, the higher incoming air relative humidity will lead to the lower outlet 270 air dry bulb temperature. For instance, at the inlet air dry bulb temperature of 30•C, the outlet air dry 271 bulb temperature were 24.8•C, 25.7•C and 26.4•C respectively for RH of 23%, 32% and 40%. This 272 shows that the proposed novel evaporative cooling system has great potential to be used in hot and dry 273 climatic conditions. On the other hand, Figure 4 shows that the outlet air dry bulb temperature tends to 274 form a linear relationship with the inlet air dry bulb temperature at the same inlet air relative humidity. The slopes of the outlet air dry bulb temperature at constant relative humidity are in the 276 range of 0.7-0.74. This means that, when increasing the inlet air temperature by 10•C, the outlet air 277 temperature will be improved by 7-7.4•C. 278 The wet bulb effectiveness of the proposed novel evaporative cooling system is illustrated in Figure 5. 288 It can be found that, with the incoming air dry bulb temperature in the range of 27-39•C, and RH of 289 23%, 32% and 40%, the wet bulb effectiveness varies from 0.32-0.45. Higher inlet air dry bulb 290 temperature leads to greater wet bulb effectiveness, due to the fact that larger temperature depression 291 is obtained for higher inlet air temperature. In addition, lower inlet air relative humidity will result in 292 higher wet bulb effectiveness. The reason is due to the fact that drier incoming air with small relative 293 humidity actually represents larger driving force of vapour pressure difference between the inlet and 294 outlet air condition. Thus the direr incoming air can potentially absorb more moisture during the water 295 evaporation process. Therefore, more latent heat will be required during the water evaporation 296 process. Consequentially, a larger amount of sensible heat of the processed air will be transferred 297 from the incoming dry air to the outgoing wet air. This leads to a much lower outlet air temperature 298 compared with air of higher incoming relative humidity. Comparable value of wet bulb effectiveness 299 obtained by Dohnal et al. [25] is also included in Figure 5. With the inlet air dry bulb temperature of 300 24.6•C, and the RH= 25%, the wet bulb effectiveness achieved by Dohnal,et al. [25] was 0.354, 301 Figure 6 illustrates the variations of dew point effectiveness with respect to different inlet air 303 conditions. When air inlet dry bulb temperature increases from 27•C to 39•C, the dew point 304 effectiveness is in the range of 0.18-0.3. At the same inlet air RH, higher inlet air temperature leads to 305 higher dew point effectiveness, due to the fact that larger temperature depression is obtained for 306 higher inlet air temperature. Moreover, when the inlet air dry bulb temperature maintains at the same 307 level, lower inlet RH will result in higher dew point effectiveness. The reason is similar to what we 308 present in last paragraph, as drier incoming air with smaller relative humidity actually represents 309 larger driving force of vapour pressure, thus leads to higher cooling performance. The dew point 310 effectiveness obtained in this experimental study is slightly better than the results achieved by Dohnal 311 et al. [25], due to the fact that the at lower air velocity, the individual fibres within the bundles were 312 shielded from the air stream, as the majority of the air will go past the outer layer of the fibre bundles. temperature, lower inlet air RH will lead to higher cooling capacity. For instance, when inlet air dry 319 bulb temperature is fixed at 30•C, the cooling capacities are 125.2W, 109.9W and 83.1W respectively 320 for the inlet air RH equals to 23%, 32% and 40%. The reason is due to the fact that, as shown in Eq. (1), 321 the cooling capacity is proportional related to the differences between the inlet air and outlet air dry 322 bulb temperature. Drier incoming air with small relative humidity actually represents larger driving 323 force of vapour pressure difference between the inlet and outlet air condition. Thus the drier incoming 324 air can potentially absorb more moisture, which leads to a much lower outlet air temperature compared 325 with air of higher incoming relative humidity. Therefore, the cooling capacity shows decreased trend 326 when increasing the inlet air RH from 23% to 32% and 40%. 327 Figure 8 shows the experimental obtained heat flux under different incoming air relative humidity. It 328 can be observed that the heat flux will increase with the improvement of incoming air Reynolds number. At the same Reynolds number, lower inlet air relative humidity will lead to greater heat flux. 330 For instance, at Reynolds number of 100, the heat flux increases from 798W to 1512W respectively 331 for RH equals to 40% and 23%. This means that approximately 1.8 times more heat is 332 transferred in the process when RH decreased from 40% to 23%. The reason is because, as 333 shown in Eq. (3), the heat flux is positively related to the inlet and outlet air dry bulb 334 temperature difference. Moreover, as shown in Figure 5, lower inlet air relative humidity will 335 yield much lower outlet air dry bulb temperature, which means greater inlet and outlet air dry bulb 336 temperature difference. This will consequentially lead to higher heat flux. Similar trend could be 337 found in Figure 9, showing the variations of mass flux with respect to Reynolds number under 338 different incoming air RH. It is obvious that higher Reynolds number will contribute to greater mass 339 flux under the same RH. While for the same Reynolds number, lower RH will lead to higher mass 340

flux. 341
Experimental determined overall heat transfer coefficient (ℎ ு ) and mass transfer coefficients (ℎ ெ ) with 342 respect to Reynolds number under different inlet air relative humidity are illustrated in Figure 10 and 343 Figure 11. An increase in Reynolds number yields better heat and mass transfer between the air stream 344 and the water inside polymer hollow fibre. Despite of different inlet air relative humidity, ℎ ு follows 345 the same linear relationship with Reynolds number, with variations less than 4.1% during the 346 experiments. The similar linear relationship between Reynolds number and ℎ ெ is illustrated in Figure  347 11. Further inspection of Figure 10 and Figure 11 indicate that, by increasing Reynolds number from 0 348 to 220, ℎ ு changes from around 60 W/m 2 K to 250 W/m 2 K (about 4.2 times), while ℎ ெ improves from 349 0.01m/s to 0. 25m/s respectively (about 2.5 times). This indicates the changes of Reynolds number has 350 more significant impact on the heat transfer coefficients than the mass transfer coefficients for the 351 proposed polymer hollow fibre integrated evaporative cooling system. 352 For the proposed novel evaporative cooling integrated hollow fibre system, the low Reynolds number 361 (less than 300) clearly shows the laminar flow regime. Therefore, the exponent α, for this study will 362 be chosen as 1/3. 363 After obtaining the heat and mass transfer coefficients hH and hM, based on Eq. (7) and Eq. (8)  achieved by Johnson et al.[15] are the worst compared with other fibre configurations. In the present 403 study, the fibre module consists of 5 fibre bundles which contain 100 fibres individually. However, the 404 mass transfer performance of such proposed fibre module shows significantly improvement compared 405 with the mass transfer conditions of fibre bundles presented by Johnson et al. [15]. The reason is due to the fact that, the spindle shape in the proposed system helps to avoid any over-shielding effect within 407 the fibre bundle. By compressing the fibre from both ends, the loosed spindle shape fibre bundle could 408 be obtained, which helps to enable better heat and mass transfer between each individual fibre and the 409 incoming air. For normal shape fibre bundle, such shielding effect is more significant at lower Reynolds 410 number, when the air will go past the outside of the fibre bundle, leaving the majority of the fibre inside 411 the bundle with very little contact with the incoming air. As the Reynolds number increases, better 412 contact between the fibres inside the bundle could be achieved, which leads to better mass transfer 413 performance, as shown in Figure 14. 414 415

Conclusions 416
A novel evaporative cooling system with hollow fiber bundles in the spindle shapes is proposed in this 417 research. With the aim to avoid the flow channelling or shielding of adjacent fibres, the fibres were 418 compressed into a spindle shape to allow maximum contact between the incoming air and the fibres. 419 This novel hollow fibre integrated evaporative cooling system will provide a comfortable indoor 420 environment for hot and dry area. Under various inlet air dry bulb temperatures (27•C, 30•C, 33•C, 421 36•C and 39•C), and various inlet air relative humidity (23%, 32% and 40%), the cooling performances 422 of the proposed novel evaporative cooling system were experimentally investigated. The variations of 423 outlet air dry bulb temperature, wet bulb effectiveness, dew point effectiveness and cooling capacity 424 were studied by varying the incoming air dry bulb temperature. Some conclusions can be found: 425 1) Increase the inlet air dry bulb temperature will lead to the increase of the outlet air dry bulb 426 temperature, cooling capacity, wet bulb effectiveness and dew point effectiveness. By keeping 427 the inlet air dry bulb temperature at constant value, increase the inlet air relative humidity will 428 lead to the decrease of cooling capacity, wet bulb effectiveness and dew point effectiveness; 429 2) The heat and mass transfer coefficients remain to be in linear relationships with respect to the 430 Reynolds number, despite of various inlet air relative humidity. With the Reynolds number in 431 the range of 10-220, the heat and mass transfer coefficients were in the range of 50-250W/m2K and 0.05-0.3m/s. Two sets of non-dimensional heat and mass transfer correlations with respect 433 to Reynolds number were deviated from the experimental results, which showed good 434 agreements with other correlations from literature; 435 3) With inlet air dry bulb temperature being the same, increase the inlet air relative humidity will 436 lead to the decrease of heat flux and mass flux. With the inlet air relative humidity being equal, 437 increase the Reynolds number of incoming air will lead to the increase of heat flux and mass the packing fraction of 0.028. In the literature, the packing fraction was 0.28 in Zhang [21], 448 which was about 10 times higher than what the hollow fiber module presented in this paper. 449 However, the cooling effectiveness was only 1.5-2 times higher than that is obtained in this 450 research. This means that the design of this hollow fiber integrated evaporative cooler prototype 451 could provide comparably cooling performance as presented by other researchers, but with 452 lower packing fraction factor(fewer fibers included). The future work could be concentrated 453 on increase the packing fraction by inserting more fibers into the one bundle, in order to 454 potentially increase the cooling effectiveness. 455 Tables   460   Table 1