- Email: [email protected]
Ceramic tubes membrane technology as a new humidification technique for gas turbine inlet air cooling
International Journal of Thermal Sciences 80 (2014) 1e10
Contents lists available at ScienceDirect
International Journal of Thermal Sciences journal homepage: www.elsevier.com/locate/ijts
Ceramic tubes membrane technology as a new humidification technique for gas turbine inlet air cooling O. Zeitoun, M. Ali*, H. Al-Ansary, A. Nuhait King Saud University, College of Engineering, Mechanical Engineering Department, P.O. Box 800, Riyadh 11421, Saudi Arabia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 March 2013 Received in revised form 21 January 2014 Accepted 22 January 2014 Available online 27 February 2014
An experimental study is conducted to cool ambient air using a new humidification technique. A wind tunnel is built with a test section comprising a matrix of ceramic tubes. These ceramic tubes are of porous design to achieve air cooling by humidification. Ambient air passes over the ceramic tubes matrix (cross flow) whereas water passes through the ceramic tubes. Air temperature and relative humidity data are measured upstream and downstream of the ceramic tube bundle used to humidify the ambient air for several air and water flow velocities. Air velocity is measured at different locations along the centerline of the rectangular wind tunnel’s cross section before the test section. Results show that the ambient temperature drops by about 10 C when the relative humidity increases from 2% to 5.4%. Heat and mass transfer analyses are made and show good agreement with correlations available in the literature. It is noticed that the evaporation process does not follow the isenthalpic lines. Therefore, heat is transferred from the air as latent and sensible heats. A 25% decrease in the duct air outlet temperature is obtained as the water velocity increased to 0.0347 m/s (9.81 107 m3/s). The results also show that the maximum estimated evaporative cooling system efficiency of the test section is about 45%. Ó 2014 Elsevier Masson SAS. All rights reserved.
Keywords: Cross flow heat exchanger Air humidification Gas turbine air compressor Heat and mass transfer
1. Introduction Cooling the air at compressor inlet is a well known technology used to increase gas turbine capacity and efficiency. Air humidification can be used to cool the air at compressor inlet. This technology is inexpensive, simple to apply and its power consumption is low. Humidification is normally carried out by spraying water in air flow upstream of the compressor inlet. Using this method requires high quality water to avoid corrosion and erosion of the compressor blades, and scale off composition on compressor blades. Furthermore, droplet drift can increase water consumption in the conventional humidification process. Membrane evaporation is a new technology used in many applications such as desalination and juice concentration. Using this technology in air humidification eliminates blade problems mentioned above and droplet drift. In addition, low quality water can be used in the humidification process. Inlet air cooling markedly enhances the performance of combustion turbines [1e6]. The turbine power increases at a lower cost per kW, and as an added benefit the heat rate also improves. Various approaches to cooling the turbine inlet
* Corresponding author. E-mail addresses: [email protected], [email protected] (M. Ali). http://dx.doi.org/10.1016/j.ijthermalsci.2014.01.019 1290-0729/Ó 2014 Elsevier Masson SAS. All rights reserved.
air have been employed. The two most common approaches (evaporative cooling and mechanical refrigeration) have been extensively applied, and are well developed and documented. Combustion turbines have ambient temperature sensitivity: both the capacity and efficiency decrease as the ambient temperature increases. The power demand of the compressor section of the turbine is proportional to the absolute temperature of the inlet air. The compressor capacity is proportional to the density of the inlet air, which is inversely proportional to the absolute temperature. Therefore higher ambient temperatures negatively affect both capacity and efficiency of the turbine. Turbine manufacturers supply curves detailing both the power output and heat rate as a function of ambient temperature. Erickson et al. [3] reported that a 300-refrigeration ton aqua ammonia refrigeration unit is required to cool the inlet of a 5 MW gas turbine from 35 C to 5 C. This cooling increases the power output by 1 MW, and the added power is at a marginal efficiency of 39%, compared to 29% for the base turbine power. Alhazmy and Najar [5] reported that the spray coolers appear to be capable of boosting the power and enhancing the efficiency of the gas turbine power plant in a way that is less expensive than cooling coils. Although the performance of spray coolers is deeply influenced by the ambient temperature and humidity, they operate efficiently during hot and dry climatic conditions. The analysis of Alhazmy and
2
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Nomenclature A ATS BTS Cp Cw d Dv F he hfg hm JD JH k km _ ev m _ m n Nu Pr Prs Q RH Re Sc Scs Sh SL
tube bank surface area, m2 after test section before test section specific heat at constant pressure, J/kg K specific heat of water, J/kg K tube diameter, m mass diffusivity of water vapor in air, m2/s correction factor evaporative heat transfer coefficient, W/m2 K latent heat of vaporization, J/kg convective mass transfer coefficient, m/s Colburn mass transfer group Colburn heat transfer group thermal conductivity, W/m K mass transfer coefficient, kg/m2 s evaporation rate, kg/s flow rate, kg/s constant Nusselt number Prandtl number Prandtl number at surface temperature heat transfer, W relative humidity Reynolds number Schmidt number Schmidt number at surface temperature Sherwood number distance between the tube bank centers parallel to the flow, m
Najar [5] have shown that the spray cooler reduces the temperature of incoming air by 3e15 C, enhancing the power by 1e7% and improving the efficiency by 3%. Membrane evaporation is a new technology which utilizes the evaporative cooling technique in air conditioning, water desalination, juice concentration and other applications [7e12]. Microporous hydrophobic membranes have been examined by Loeb [7] for possible use as containers in the evaporative cooling of water, particularly in desert climates. An experimental determination was made of the overall heat and mass transfer coefficients of these membranes while surmounting contained water and with air flowing over the surface of the membranes. Recently, Zhang [13] has reported a numerical and experimental study about parallelplates membrane cores used in air-to-air heat exchangers for fresh air heat and moisture recovery. His results indicated that for these membrane structures, when the channel pitch is below 2 mm, the flow distribution is quite homogeneous and the sensible and latent heat performance deteriorations due to flow maldistribution are below 9% and can be neglected. However, when the channel pitch is larger than 2 mm, the maldistribution is quite large and the consequent thermal and latent performance can deteriorate by 28%. More recently, a numerical simulation for mass transfer through a porous membrane of parallel straight channels has been reported by Lu and Lu [14]. In their study, two types of flows, channel flow and ultra-filtration flow, were physically described. Their results displayed the flow and solute distribution patterns inside channels, described the ultra-filtration profiles along the surface of the porous membrane and disclosed an existent nanoscale reverse osmosis problem. Ceramics and ceramic matrix composites for heat exchangers in advanced thermal systems have
ST T umax v
distance between the tube bank centers normal to the flow, m temperature, C water maximum velocity through the tube bank, m/s velocity, m/s
Greek symbols pressure loss, Pa absolute viscosity, Pa s density, kg/m3 average air density, kg/m3 humidity ratio, kg (vapor)/kg (dry air) average humidity ratio, kg (vapor)/kg (dry air) humidity ratio at the tube surface, kg (vapor)/kg (dry air)
DP m r ram u um us
Subscripts a air d dry di dry bulb at inlet do dry bulb at outlet ew outlet water i inlet iw inlet water m mean o outlet s at the tube surface w water wb wet bulb win wet bulb at inlet
been reviewed by Sommers et al. [15]. In their paper, the current state-of-the-art of ceramic materials for use in a variety of heat transfer systems was reported. It should also be mentioned that other methods of cooling the inlet air are known as wet media evaporative cooling technology, which offer 85e90% evaporation efficiency, and may not require high water quality but they need huge amounts of water. These methods offer reduced risk of erosion to the compressor blades and corrosion to turbine inlet duct structure. It should be noted that the current suggested method presents a new interesting engineering concept to enhance the performance of the gas turbine in spite of its low efficiency. Therefore, further research needs to be performed to improve the efficiency of the proposed technology, such as increasing the number of tubes and reducing their diameters, which leads to increase the exchange surface between the ceramic tubes and air flow. Another advantage is the possibility of using recycled water (chemically treated water) since the fresh water in arid areas like Saudi Arabia is mostly available through desalination plants and commonly used for human consumption. In this paper, evaporation technology is used to humidify the air for cooling before it enters to the compressor of a gas turbine. A wind tunnel is built and a matrix of ceramic tubes is used as a test section where water passes through in cross flow configuration. The relative humidity, pressure loss, air velocity and water velocity are measured before and after the test section. Heat and mass transfer analyses are made. Results show that the proposed technology has an efficiency not to exceed 45%. Therefore, this study can be considered of preliminary nature and more detailed technical study including economic aspects will need to be examined for the proposed method to be implemented in real life applications.
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Fig. 1. Showing the conical adaptor attached to a 2.2 kW fan (first part of the wind tunnel).
2. Experimental setup The experimental apparatus consists of a blower, wind tunnel and a matrix of ceramic tubes serving as a test section. The wind tunnel has three parts. The first part is a conical adaptor of length
3
150 cm that converts the blower circular section of 50 cm in diameter to a rectangular section of 28 cm 14 cm as seen in Fig. 1. The blower part has a fan of capacity 2.2 kW (1710 RPM) as seen in Fig. 1 and schematically in Fig. 2(a). The complete wind tunnel assembly is shown in Fig. 2(b). The second part is the test section and the third part is the settling or the straightener section. This settling section consists of three rectangular ducts each of which is 102 cm long with a cross section of 28 cm 14 cm in addition to one more duct of 50 cm length with the same cross section that makes the total settling section 356 cm long (Fig. 2(a)). A form of honeycomb is used at the entrance section to provide more uniform flow. The second part is the test section which consists of 39 ceramic tubes (material: membrane element made of Al2O3, 1/6 monochannel- design, 0.2 mm pore diameter, 300 mm length, as specified by the manufacturer) arranged as a matrix of cross-flow heat exchanger as seen in Fig. 3. The outer and inner diameter of each tube of the matrix is 10.0 mm and 6.0 mm respectively. Fig. 3 shows two compartments for water inlet and outlet. Fig. 4(a)e(c) show schematically the dimensions of the test section in cm. It should be noted that, the ceramic porous tubes, as supplied by the manufacturer, have mean pore diameters of 0.2 mm. Those pores are very wide to be used for humidification of the air with water in a vapor form. Therefore, a
Fig. 2. Schematic of the wind tunnel, (a) wind tunnel (dimensions in cm); 1 e test section, 2 e conical adaptor, 3 e variable speed fan, 4 e intake duct, 5 e inlet air and (b) complete wind tunnel assembly.
4
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Fig. 5. Locations of the velocity sensors measurements (dimensions are in mm).
Fig. 3. The test section made of Plexiglas showing the ceramic tubes and the two compartments for inlet and outlet of water.
special experiment was conducted to reduce the pore size by pumping flow of alumina nanofluid with 20% concentration by weight through the ceramic tubes. The porosity of the modified tubes was tested against that of the original ones for fixed amount of water in still air. The results show that the modified tubes have an evaporation rate of 2.717 10-6 l/s, whereas the original tubes have 1.625 104 l/s. The measurement devices such as digital
Fig. 4. Schematic of the test section, (a) test section dimensions, (b) water feed to the test section and (c) staggered membrane tubes test section. All dimensions are in cm.
low range water flow meters, velocity and temperature sensors, pressure gauges and relative humidity devices are installed on the wind tunnel and connected to data acquisition systems. Relative humidity measuring devices and temperature sensors are installed in their housing on the duct before and after the test section. The wind tunnel is equipped with a variable speed fan to achieve variable air flow velocities. Two computers are used; one is connected to the data acquisition system for the air temperature and relative humidity measurements and the other for air velocity measurements. Air velocity is measured using an 8-channel hot wire anemometer. 3. Experimental procedure Control of air flow is established using a variable speed motor controller by adjusting the power source frequency which is fed to the fan motor. Water flow rate is adjusted by a water control valve. Air velocity is measured by hot wire anemometers. Seven velocity sensors are fixed on vertical strips at vertical y-distances 7.42, 28.4, 51.24, 70.0, 88.76, 111.58 and 132.58 mm to scan the air velocity. The vertical strip holding the sensors slides along the horizontal direction of the rectangular cross section to scan the velocity in the horizontal direction. The coordinates of the sensors are shown in Fig. 5. Following this procedure, air velocities are measured at 42 points over the entire cross section. A data acquisition system is connected to a PC and used to collect air velocities. The mean air velocity is estimated based on the 42 measured values. Water flow rate through the test section is measured using a turbine flow meter. Temperature along the air duct is measured using K-type thermocouples and thermistors. Four thermocouples are used, two of them at the inlet and the others at the outlet to measure the dry and wet bulb temperatures. Two thermocouples are fixed before and after the test section to measure the air temperatures across the test section. Two more thermocouples are used to measure the water temperatures in the water tank and at the outlet of the test section. One thermocouple is used to measure the room temperature. Humidity sensors are used to measure the relative humidity before and after the test section. These sensors are also equipped with thermistors to measure temperatures. Two pressure transducers are used to measure the pressure drop across the test section. Another data acquisition system is connected to a laptop to collect the data of humidity sensors, pressure transducers and thermocouples. Table 1 shows the uncertainty in the measured parameters. The thermocouples and data acquisition system are calibrated at ice and boiling points. At the ice point, the error in thermocouple readings is in the range 0.2 to 0.1 C. However, the error at the boiling point is in the range 0.15e0.4 C. The used flow meter is a turbine type (FTB602) made by Omega Engineering, INC. The flow meter is
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Consequently, the data can be represented in terms of Colburn heat transfer group JH (Zhukauskas [16]):
Table 1 Uncertainty of the measured parameters. Temperature Relative humidity Pressure Velocity Water flow rate
5
0.2 to þ0.4 C 2% R.H 0.25% 2% 2%
JH ¼
Nu Re Pr 1n
(8)
Similarity between heat and mass transfer leads to (Zhukauskas [16]):
calibrated using a balance scale and a stop watch. The difference between the reading of the flow meter and that of the measured flow rates is 2% at most.
JD ¼
Sh Re Sc1n
(9)
4. Heat and mass transfer analysis In this section, an attempt is made to analyze the heat and mass transfer between the ceramic tubes and the flowing air. The heat transferred from the air through the membrane tube bank is used to evaporate the water through the membrane surface. Therefore, the heat lost from the air due to water evaporation is given by:
_ a Cp ðTdi Tdo Þ Qw Qa ¼ m
(1)
where Tdi and Tdo are air temperatures at test section inlet and outlet respectively, and Qw is the heat gained by the water flow inside the membrane tube:
_ w Cw ðTiw Tew Þ Qw ¼ m
(2)
where Tiw and Tew are water temperatures at inlet and outlet respectively. The evaporation heat transfer coefficient is estimated from:
he ¼
Qa AðTm Twb Þ
(3)
where A is the tube bank surface area, Twb is the wet bulb temperature which represents the film temperature on the membrane surface and Tm is the mean temperature defined by:
Tm ¼
Tdi þ Tdo 2
(4)
Nusselt number is defined by:
Nu ¼
he d k
(5)
where d is tube diameter and k is the air thermal conductivity. The Reynolds number based on the maximum velocity through the tube bank is calculated from:
Re ¼
rUmax d m
(6)
The general average correlation for the entire tube bundle has been reported by Zhukauskas [16] of the form
0:2 S Nu ¼ F 0:35 T Re0:6 Pr n ðPr=Prs Þ0:25 SL
(7)
This equation is valid for 1000 Re 200,000 and ST and SL are the distances between tube centers normal and parallel to the flow through the tube bank, respectively. F is a correction factor for banks with less than 16 rows and n ¼ 0.36. For the current experiment: F ¼ 0.95 corresponding to 7 rows and (ST/SL) ¼ 1.
Fig. 6. Temperature profiles for 1 h span of air at the inlet and exit from the duct and at the inlet and exit from the test section at fixed air velocity and for various water velocity; (a) vw ¼ 0.01667 m/s, (b) vw ¼ 0.0287 m/s, and (c) test section inlet and outlet temperature at three different water velocities.
6
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Table 2 Dry and wet bulb temperatures at the inlet and outlet of both the test section and the duct where the evaporative cooling system efficiency estimated for the test section. No Duct inlet
Test section inlet
Test section outlet
Duct exit
Test section efficiency %
Td, C Twb, C Td, C Twb, C Td, C Twb, C Td, C Twb, C 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
36.68 37.47 38.5 39.98 39.59 39.85 42.2 43.18 35.06 39.35 40.76 41.19 35.5 39.15 39.81 40.64 35.76 37.3 36.66 40.38 40.58
15 15.6 15.5 17.5 17.5 16.8 18.5 18.7 16 16 18.5 20 16.3 18 17 18.5 16 16.6 14.5 18.2 17.5
35.86 36.79 37.81 38.52 38.64 39.21 41.61 42.49 34.7 38.82 39 40.63 35.21 38.56 38.96 41.06 35.65 36.96 36.64 40.5 40.68
14.71 15.36 15.26 17.02 17.19 16.59 18.31 18.48 15.88 15.82 17.94 19.83 16.2 17.81 16.72 18.63 15.96 16.49 14.49 18.24 17.53
29.99 30.75 31.04 31.58 29.14 32.33 35 36 29.5 33 32.5 32 26.66 32.26 32.08 35.92 31.8 32.69 33.2 35.25 34.39
13.18 14.36 14.77 16.57 15.4 15.01 17.92 17.39 15.24 16.93 15.95 17.65 12.87 16.05 14.44 17.82 13.97 16.58 13.7 18.04 17.44
29.51 30.59 31.67 31.36 28.02 29.56 32.19 33.31 28.83 30.26 29.8 30.03 27.77 29.26 29.54 34.01 31.88 29.85 28.81 30.55 30.1
13 14.3 15 16.5 15 14 17 16.5 15 16 15 17 13.3 15 13.5 17.2 14 15.6 12 16.5 16
27.75 28.18 30.02 32.28 44.29 30.42 28.37 27.03 27.63 25.30 30.86 41.49 44.98 30.36 30.94 22.92 19.55 20.86 15.53 23.58 27.17
where JD is the Colburn mass transfer group and Sh is the Sherwood number:
Sh ¼
hm d Dv
hm ¼ km =ram
(17)
where ram is the average air density. 5. Results and discussion Measurements of the temperature and relative humidity are taken for the ceramic tube test section shown in Fig. 3 at various air and water velocities. Fig. 6(a, b) shows the temperature profiles for two different water flow velocities of 0.0167 m/s (4.71 107 m3/s) in Fig. 6(a) and 0.0287 m/s (8.11 107 m3/s) in Fig. 6(b) as the time goes for 1 h for fixed air velocity of 3.52 m/s (0.138 m3/s). This figure shows that there is minimum heat leakage from the wall of the wind tunnel since there is not much difference between the temperatures at the duct inlet and the test section inlet. The same can be said about the temperature between the test section outlet and the duct outlet as an average. It should be noted that in Fig. 6(b), at the beginning of the experiment, the outlet temperature of the duct is higher than that of the test section by about 5 C. However when the experiment reaches steady state at time between 15 and 35 min the two temperatures become equal. The reason for initial discrepancy could be that the sensor at the outlet of the duct is affected initially by the lab temperature which is higher than that of the outlet test section temperature. Furthermore, on average, there is about 6 C cooling in the air at the exit from the duct. This temperature drop represents a cooling ratio of about 17% compared to inlet air to the duct which based on an average of 35 C inlet air temperature (Fig. 6(a)). On the other hand, there is a 10 C drop in
(10)
where hm is the convective mass transfer coefficient and Dv is the mass diffusivity of water vapor in air. The Schmidt number Sc is:
Sc ¼
m rDv
(11)
According to the similarity, Sh number can be presented by (Zhukauskas [16]):
Sh ¼
0:25 1n 0:2 Sc S Sc F 0:35 T Re0:6 Sc0:36 Pr Scs SL
(12)
The evaporation rate at tube bank surface can be estimated from:
_ ev ¼ Qa =hfg m
(13)
where hfg is the water latent heat of vaporization and the humidity ratio increase can be obtained from:
_ ev =m _a uo ui ¼ m
(14)
The mass transfer coefficient can be estimated from:
km ¼
_ ev m Aðus um Þ
(15)
where us is the humidity ratio at the tube surface and is obtained from the psychometric chart at the measured wet bulb temperature and um is the average humidity ratio:
um ¼ ðue þ ui Þ=2
(16)
According to ASHRAE [17], the convective mass transfer coefficient hm is related to the mass transfer coefficient km by:
Fig. 7. Effect of water flow rate on relative humidity at two different ambient conditions and for various inlet water velocities; (a) vw ¼ 0.01667 m/s and (b) vw ¼ 0.0287 m/s.
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
7
Fig. 8. Temperature profiles for 1 h span of air at the inlet and exit from the duct and at the inlet and exit from the test section at fixed air velocity; (a) for water velocity 0.0236 m/s and (b) for 0.0347 m/s.
Fig. 9. Effect of water velocity on the relative humidity at two different ambient conditions and for fixed air velocity for; (a) vw ¼ 0.0236 m/s and (b) vw ¼ 0.0347 m/s.
the air leaving the duct as seen in Fig. 6(b) which represents a 25% cooling ratio of the inlet temperature (based on about an average 40 C inlet air temperature). Fig. 6(c) shows the inlet and outlet temperature profiles of the test section at 3.61 m/s air velocity and for three water velocities 0.0167 m/s, 0.0229 m/s and 0.0343 m/s. The average temperature difference drop between the inlet and outlet are 9.5 C, 6.8 C and 6.49 C for the different velocities respectively. It should be noted that the numerical values of the average temperature at the inlet and outlet of the test section can be obtained from Table 2, runs number 5, 6 and 8 corresponding to various water velocities respectively. This figure indicates also that as the water velocity increases the temperature drop between the inlet and outlet decreases as seen for run 5 and 6 (Table 2) for almost the same inlet temperature. Fig. 7 shows the relative humidity for fixed air flow velocity and for several water velocities at various inlet air conditions. The inlet relative humidity starts (at time ¼ 0) at about 9.5% at the inlet to the test section and increases to about 13% at the outlet of the test section as shown in Fig. 7(a). It should be noted that as the time goes during the 1 h experiment both the inlet and outlet relative humidity decrease keeping the difference almost constant as seen in Fig. 7(a). However, for the other ambient conditions, the relative humidity starts at about 2% and increases to about 5.4% at the exit of the test section as seen in Fig. 7(b). This observation confirms the decrease in the air temperature seen in Fig. 6(a, b). Fig. 8(a, b) shows the temperature profiles for suction air velocity 4.26 m/s (0.167 m3/s) and for two water velocities 0.0236 m/s (6.67 107 m3/s) in Fig. 8(a) and 0.0347 m/s (9.81 10 m3/s) in Fig. 8(b). Fig. 8(a) shows that the
ambient inlet duct temperature is almost constant at an average value of 40.38 C. This figure also shows that the average duct outlet temperature is 30.55 C with a drop in air temperature of 9.83 C which represents a 24.3% cooling ratio of the ambient temperature. It is worth mentioning that the duct outlet temperature is lower than the test section outlet temperature, which could be attributed to some small droplets of water that evaporate in the passage between the test section outlet and the duct outlet. A 25.8% decrease in the duct air outlet temperature is noticed in Fig. 8(b) as the water velocity increased to 0.0347 m/s (9.81 107 m3/s). The corresponding relative humidity is shown in Fig. 9(a, b). The average relative humidity increases from about 0.2% to 2.3% as seen in Fig. 9(a and b). Water temperature at the inlet and outlet of the test section is shown in Fig. 10(a) and (b) for two different water velocities and at fixed air speed of 3.52 m/s (0.138 m3/s). As seen in Fig. 10(a) at water velocity 0.01667 m/s (4.71 107 m3/s), the maximum temperature difference between the inlet and outlet is about 2 C. However, no remarkable difference is observed as the water velocity increased to 0.0249 m/s (7.04 107 m3/s) as seen in Fig. 10(b). It should be noted that data in Fig. 10(a) are corresponded to the data given in Figs. 6(a) and 7(a) for temperature and relative humidity respectively. Therefore, Fig. 10 shows that part of the absorbed heat from the air is used as a latent heat for evaporating the water as discussed earlier in the heat and mass transfer analysis section. It should be mentioned that low water flow rates are used such that the water can go through the pores in a vapor form to avoid droplets or film of water at the outside surface of the tubes. On the other hand, it is noticed that the evaporation process on the
8
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Fig. 10. Effect of water velocity on the water temperature before and after the test section at fixed air velocity; (a) vw ¼ 0.01667 m/s and (b) vw ¼ 0.0249 m/s.
psychometric chart does not follow isenthalpic lines. This means that not all the dissipated energy from the air is transferred as latent heat to the evaporating water but part of it is used as a sensible heat. Samples of the psychometric chart processes are shown in Fig. 11(a, b, c). These plots are based on dry and wet bulb temperatures at the test section inlet and outlet. Fig. 11(a, b, c) corresponds to some runs shown in Table 2. It is clear from these figures that the process curves do not exactly follow the isenthalpic lines; therefore some significant non-evaporative heat transfer is present. Fig. 12 summarizes the relation between the water flow rates through the ceramic pipes for different air velocities and the relative humidity before (BTS) and after the test section (ATS). This figure shows that the relative humidity decreases as the water flow rate increases at higher values of air velocities. This figure also summarizes the rate of increase of the relative humidity RH before (BTS) and after the test section (ATS) as explained earlier in Fig. 7(a,b) and 9 (a,b) for some water and air velocities. Table 2 shows the performance of the experimental data at the test section using the evaporative cooling system efficiency (Tdin Tdout)/(Tdin Twin). It should be noted that the wet bulb temperature at the duct inlet and outlet are measured using sensors covered with wet wick. Wet bulb temperatures at test section inlet and outlet are obtained from the psychometric chart (EES software) assuming that the process from the duct inlet to test section inlet has the same humidity ratio, and that the process from the test section outlet to duct outlet has a constant humidity ratio. Table 3 shows the air pressure losses for different air and water velocities along the test section. As seen in this Table as the air velocity increases the pressure loss increases up to a maximum value of 44 Pa.
Fig. 11. Psychometric chart for some selected processes; numbers in the figure referred to the experimental runs shown in Table 2; (a) runs 1, 2 and 4 (b) runs 5, 6 and 8 and (c) runs 18, 19 and 20.
Comparison between the staggered tube bank heat transfer represented by the Colburn heat transfer group JH Eq. (8) and the current experimental data is shown in Fig. 13. The present experimental data show good agreement with Eq. (8) since they lie within 20%. Furthermore, comparison between the predicted Colburn mass transfer group JD Eq. (9) and the current experimental data is shown in Fig. 14. Again our experimental data is within 20% of correlation Eq. (9). 6. Conclusions Results show that using the modified ceramic tubes enhances the heat transfer between the air and water vapor. This enhancement is reflected by a maximum and minimum temperature drop of 11 C and 4 C of the dry air respectively. It is noticed that some distance should be left between the test section outlet and the duct outlet to give a chance for any droplet of water already in the air to evaporate before exiting the duct for further use. As expected, the relative humidity downstream of the test section is found to increase by about 3.5% compared to ambient conditions. Heat and mass transfer analysis for the tested ceramic tubes arrangement compared within 20% with the published (8, 9) correlations for
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
9
Fig. 14. Comparison of the current experimental data with the Colburn mass transfer group (9) for staggered tube bank.
Fig. 12. The effect of water flow rate on the relative humidity before and after the test section for various air velocities.
Table 3 Air pressure losses at various values of air and water at the test section. No.
Vw m/s
Va m/s
DP Pa
1 2 3 4 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
0.0167 0.0226 0.0287 0.0344 0.0229 0.0287 0.0343 0.0169 0.0231 0.0286 0.0340 0.0170 0.0225 0.0280 0.0343 0.0330 0.0406 0.0163 0.0236 0.0347
3.5188 3.5188 3.5188 3.5188 3.6145 3.6145 3.6145 4.0334 4.0334 4.0334 4.0334 3.8031 3.8031 3.8031 3.8031 4.2575 4.2575 4.2575 4.2575 4.2575
7.79 22.15 10.33 13.35 10.13 10.60 10.96 22.06 17.33 23.99 28.21 8.36 12.36 18.60 16.80 17.85 25.57 45.73 36.47 44.00
Fig. 13. Comparison of the current experimental data with the Colburn heat transfer group JH (8) for staggered tube bank.
the staggered tubes arrangement. It is noticed that the evaporation process does not follow the isenthalpic lines; therefore, heat is transferred from the air as latent and sensible heats. The experimental data show that ceramic tube membrane technology could be implemented for gas turbine inlet air cooling without concerns of compressor blades erosion due to water droplets associated with water spray technology (also known as fogging technology). However, more research needs to be performed to improve the efficiency of the proposed technology. Acknowledgments The authors would like to acknowledge the support of National Plan for Science, Technology, & Innovation (NPST) at King Saud University under project number ENE220-02-08. This support is highly appreciated and acknowledged. References [1] N. Ravi Kumar, K. Rama Krishna, A.V.S. Rama Raju, Improved gas turbine efficiency using spray cooler and alternative regenerator configuration, in: ICRAMME 05, Proceedings of the International Conference on Recent Advances in Mechanical &Materials Engineering, 30e31 May 2005. Kuala Lumpur, Malaysia, Paper No. 10. [2] S. Jolly, J. Nitzken, D. Shepherd, Direct Spray System for Inlet Air Cooling W 501 B5, Presented at the Power-Gen International, Dallas, Texas, December 9e 11, 1998. [3] D.C. Erickson, I. Kyung, G. Anand, E.E. Makar, Aqua absorption turbine inlet cooler, in: Proceedings of ASME International Mechanical Engineering Congress & Exposition Washington, D.C, November 16e21, 2003. [4] D.V. Punwani, R. Pasteris, GT inlet-air cooling boosts output on warm days to increase revenue, Comb. Cycle J. Fourth Quarter (2003) 1e3. [5] M.M. Alhazmy, Y.S.H. Najjar, Augmentation of gas turbine performance using air coolers, Appl. Therm. Eng. 24 (2004) 415e429. [6] M. Ameri, S.H. Hejazi, The study of capacity enhancement of the Chabahar gas turbine installation using an absorption chiller, Appl. Therm. Eng. 24 (2004) 59e68. [7] S. Loeb, Membrane evaporative cooling to 30 C or less, “Membrane evaporative cooling of contained water, Ann. N. Y. Acad. Sci. 984 (March 2003) 515e 527. Advanced Membrane Technology. [8] D.W. Johnson, C. Yavuzturk, J. Pruis, Analysis of heat and mass transfer phenomena in hollow fiber membranes used for evaporative cooling, J. Membr. Sci. 227 (1e2) (2003) 159e171. [9] D. Bessarabov, Z. Twardowski, New opportunities for osmotic membrane distillation, Membr. Technol. (7) (2006) 7e11. [10] E. Drioli, V. Calabrd, Y. Wu, Microporous membranes in membrane distillation, Pure Appl. Chem. 58 (12) (1986) 1657e1662. [11] F. Tillberg, ZLD-Systems An Overview, Department of Energy Technology, Royal Institute of Technology, KTH, Stockholm, 2004. [12] H.J. Zwijnenberg, G.H. Koops, M. Wessling, Solar driven membrane evaporation for desalination processes, J. Membr. Sci. 250 (2005) 235e246. [13] L. Zhang, Performance deteriorations from flow maldistribuion in air-to-air heat exchangers: a parallel-plates membrane core case, Numer. Heat Transfer Part A 56 (2009) 746e763.
10
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
[14] J. Lu, W. Lu, A numerical simulation for mass transfer through the porous membrane of parallel straight channels, Int. J. Heat Mass Transfer 53 (2010) 2404e2413. [15] A. Sommers, Q. Wang, X. Han, C. T’Joen, Y. Park, A. Jacobi, Ceramics and ceramic matrix composites for heat exchangers in advanced thermal systemsA review, Appl. Therm. Eng. 30 (2010) 1277e1291.
[16] A. Zhukauskas, Heat transfer from tubes in cross flow, in: J.P. Hartnett, T.F. Irvine Jr. (Eds.), Advances in Heat Transfer, vol. 8Academic Press, New York, 1972. [17] ASHRAE, Chapter 6 Mass Transfer, ASHRAE Handbook e Fundamental, 2009.
Contents lists available at ScienceDirect
International Journal of Thermal Sciences journal homepage: www.elsevier.com/locate/ijts
Ceramic tubes membrane technology as a new humidification technique for gas turbine inlet air cooling O. Zeitoun, M. Ali*, H. Al-Ansary, A. Nuhait King Saud University, College of Engineering, Mechanical Engineering Department, P.O. Box 800, Riyadh 11421, Saudi Arabia
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 March 2013 Received in revised form 21 January 2014 Accepted 22 January 2014 Available online 27 February 2014
An experimental study is conducted to cool ambient air using a new humidification technique. A wind tunnel is built with a test section comprising a matrix of ceramic tubes. These ceramic tubes are of porous design to achieve air cooling by humidification. Ambient air passes over the ceramic tubes matrix (cross flow) whereas water passes through the ceramic tubes. Air temperature and relative humidity data are measured upstream and downstream of the ceramic tube bundle used to humidify the ambient air for several air and water flow velocities. Air velocity is measured at different locations along the centerline of the rectangular wind tunnel’s cross section before the test section. Results show that the ambient temperature drops by about 10 C when the relative humidity increases from 2% to 5.4%. Heat and mass transfer analyses are made and show good agreement with correlations available in the literature. It is noticed that the evaporation process does not follow the isenthalpic lines. Therefore, heat is transferred from the air as latent and sensible heats. A 25% decrease in the duct air outlet temperature is obtained as the water velocity increased to 0.0347 m/s (9.81 107 m3/s). The results also show that the maximum estimated evaporative cooling system efficiency of the test section is about 45%. Ó 2014 Elsevier Masson SAS. All rights reserved.
Keywords: Cross flow heat exchanger Air humidification Gas turbine air compressor Heat and mass transfer
1. Introduction Cooling the air at compressor inlet is a well known technology used to increase gas turbine capacity and efficiency. Air humidification can be used to cool the air at compressor inlet. This technology is inexpensive, simple to apply and its power consumption is low. Humidification is normally carried out by spraying water in air flow upstream of the compressor inlet. Using this method requires high quality water to avoid corrosion and erosion of the compressor blades, and scale off composition on compressor blades. Furthermore, droplet drift can increase water consumption in the conventional humidification process. Membrane evaporation is a new technology used in many applications such as desalination and juice concentration. Using this technology in air humidification eliminates blade problems mentioned above and droplet drift. In addition, low quality water can be used in the humidification process. Inlet air cooling markedly enhances the performance of combustion turbines [1e6]. The turbine power increases at a lower cost per kW, and as an added benefit the heat rate also improves. Various approaches to cooling the turbine inlet
* Corresponding author. E-mail addresses: [email protected], [email protected] (M. Ali). http://dx.doi.org/10.1016/j.ijthermalsci.2014.01.019 1290-0729/Ó 2014 Elsevier Masson SAS. All rights reserved.
air have been employed. The two most common approaches (evaporative cooling and mechanical refrigeration) have been extensively applied, and are well developed and documented. Combustion turbines have ambient temperature sensitivity: both the capacity and efficiency decrease as the ambient temperature increases. The power demand of the compressor section of the turbine is proportional to the absolute temperature of the inlet air. The compressor capacity is proportional to the density of the inlet air, which is inversely proportional to the absolute temperature. Therefore higher ambient temperatures negatively affect both capacity and efficiency of the turbine. Turbine manufacturers supply curves detailing both the power output and heat rate as a function of ambient temperature. Erickson et al. [3] reported that a 300-refrigeration ton aqua ammonia refrigeration unit is required to cool the inlet of a 5 MW gas turbine from 35 C to 5 C. This cooling increases the power output by 1 MW, and the added power is at a marginal efficiency of 39%, compared to 29% for the base turbine power. Alhazmy and Najar [5] reported that the spray coolers appear to be capable of boosting the power and enhancing the efficiency of the gas turbine power plant in a way that is less expensive than cooling coils. Although the performance of spray coolers is deeply influenced by the ambient temperature and humidity, they operate efficiently during hot and dry climatic conditions. The analysis of Alhazmy and
2
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Nomenclature A ATS BTS Cp Cw d Dv F he hfg hm JD JH k km _ ev m _ m n Nu Pr Prs Q RH Re Sc Scs Sh SL
tube bank surface area, m2 after test section before test section specific heat at constant pressure, J/kg K specific heat of water, J/kg K tube diameter, m mass diffusivity of water vapor in air, m2/s correction factor evaporative heat transfer coefficient, W/m2 K latent heat of vaporization, J/kg convective mass transfer coefficient, m/s Colburn mass transfer group Colburn heat transfer group thermal conductivity, W/m K mass transfer coefficient, kg/m2 s evaporation rate, kg/s flow rate, kg/s constant Nusselt number Prandtl number Prandtl number at surface temperature heat transfer, W relative humidity Reynolds number Schmidt number Schmidt number at surface temperature Sherwood number distance between the tube bank centers parallel to the flow, m
Najar [5] have shown that the spray cooler reduces the temperature of incoming air by 3e15 C, enhancing the power by 1e7% and improving the efficiency by 3%. Membrane evaporation is a new technology which utilizes the evaporative cooling technique in air conditioning, water desalination, juice concentration and other applications [7e12]. Microporous hydrophobic membranes have been examined by Loeb [7] for possible use as containers in the evaporative cooling of water, particularly in desert climates. An experimental determination was made of the overall heat and mass transfer coefficients of these membranes while surmounting contained water and with air flowing over the surface of the membranes. Recently, Zhang [13] has reported a numerical and experimental study about parallelplates membrane cores used in air-to-air heat exchangers for fresh air heat and moisture recovery. His results indicated that for these membrane structures, when the channel pitch is below 2 mm, the flow distribution is quite homogeneous and the sensible and latent heat performance deteriorations due to flow maldistribution are below 9% and can be neglected. However, when the channel pitch is larger than 2 mm, the maldistribution is quite large and the consequent thermal and latent performance can deteriorate by 28%. More recently, a numerical simulation for mass transfer through a porous membrane of parallel straight channels has been reported by Lu and Lu [14]. In their study, two types of flows, channel flow and ultra-filtration flow, were physically described. Their results displayed the flow and solute distribution patterns inside channels, described the ultra-filtration profiles along the surface of the porous membrane and disclosed an existent nanoscale reverse osmosis problem. Ceramics and ceramic matrix composites for heat exchangers in advanced thermal systems have
ST T umax v
distance between the tube bank centers normal to the flow, m temperature, C water maximum velocity through the tube bank, m/s velocity, m/s
Greek symbols pressure loss, Pa absolute viscosity, Pa s density, kg/m3 average air density, kg/m3 humidity ratio, kg (vapor)/kg (dry air) average humidity ratio, kg (vapor)/kg (dry air) humidity ratio at the tube surface, kg (vapor)/kg (dry air)
DP m r ram u um us
Subscripts a air d dry di dry bulb at inlet do dry bulb at outlet ew outlet water i inlet iw inlet water m mean o outlet s at the tube surface w water wb wet bulb win wet bulb at inlet
been reviewed by Sommers et al. [15]. In their paper, the current state-of-the-art of ceramic materials for use in a variety of heat transfer systems was reported. It should also be mentioned that other methods of cooling the inlet air are known as wet media evaporative cooling technology, which offer 85e90% evaporation efficiency, and may not require high water quality but they need huge amounts of water. These methods offer reduced risk of erosion to the compressor blades and corrosion to turbine inlet duct structure. It should be noted that the current suggested method presents a new interesting engineering concept to enhance the performance of the gas turbine in spite of its low efficiency. Therefore, further research needs to be performed to improve the efficiency of the proposed technology, such as increasing the number of tubes and reducing their diameters, which leads to increase the exchange surface between the ceramic tubes and air flow. Another advantage is the possibility of using recycled water (chemically treated water) since the fresh water in arid areas like Saudi Arabia is mostly available through desalination plants and commonly used for human consumption. In this paper, evaporation technology is used to humidify the air for cooling before it enters to the compressor of a gas turbine. A wind tunnel is built and a matrix of ceramic tubes is used as a test section where water passes through in cross flow configuration. The relative humidity, pressure loss, air velocity and water velocity are measured before and after the test section. Heat and mass transfer analyses are made. Results show that the proposed technology has an efficiency not to exceed 45%. Therefore, this study can be considered of preliminary nature and more detailed technical study including economic aspects will need to be examined for the proposed method to be implemented in real life applications.
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Fig. 1. Showing the conical adaptor attached to a 2.2 kW fan (first part of the wind tunnel).
2. Experimental setup The experimental apparatus consists of a blower, wind tunnel and a matrix of ceramic tubes serving as a test section. The wind tunnel has three parts. The first part is a conical adaptor of length
3
150 cm that converts the blower circular section of 50 cm in diameter to a rectangular section of 28 cm 14 cm as seen in Fig. 1. The blower part has a fan of capacity 2.2 kW (1710 RPM) as seen in Fig. 1 and schematically in Fig. 2(a). The complete wind tunnel assembly is shown in Fig. 2(b). The second part is the test section and the third part is the settling or the straightener section. This settling section consists of three rectangular ducts each of which is 102 cm long with a cross section of 28 cm 14 cm in addition to one more duct of 50 cm length with the same cross section that makes the total settling section 356 cm long (Fig. 2(a)). A form of honeycomb is used at the entrance section to provide more uniform flow. The second part is the test section which consists of 39 ceramic tubes (material: membrane element made of Al2O3, 1/6 monochannel- design, 0.2 mm pore diameter, 300 mm length, as specified by the manufacturer) arranged as a matrix of cross-flow heat exchanger as seen in Fig. 3. The outer and inner diameter of each tube of the matrix is 10.0 mm and 6.0 mm respectively. Fig. 3 shows two compartments for water inlet and outlet. Fig. 4(a)e(c) show schematically the dimensions of the test section in cm. It should be noted that, the ceramic porous tubes, as supplied by the manufacturer, have mean pore diameters of 0.2 mm. Those pores are very wide to be used for humidification of the air with water in a vapor form. Therefore, a
Fig. 2. Schematic of the wind tunnel, (a) wind tunnel (dimensions in cm); 1 e test section, 2 e conical adaptor, 3 e variable speed fan, 4 e intake duct, 5 e inlet air and (b) complete wind tunnel assembly.
4
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Fig. 5. Locations of the velocity sensors measurements (dimensions are in mm).
Fig. 3. The test section made of Plexiglas showing the ceramic tubes and the two compartments for inlet and outlet of water.
special experiment was conducted to reduce the pore size by pumping flow of alumina nanofluid with 20% concentration by weight through the ceramic tubes. The porosity of the modified tubes was tested against that of the original ones for fixed amount of water in still air. The results show that the modified tubes have an evaporation rate of 2.717 10-6 l/s, whereas the original tubes have 1.625 104 l/s. The measurement devices such as digital
Fig. 4. Schematic of the test section, (a) test section dimensions, (b) water feed to the test section and (c) staggered membrane tubes test section. All dimensions are in cm.
low range water flow meters, velocity and temperature sensors, pressure gauges and relative humidity devices are installed on the wind tunnel and connected to data acquisition systems. Relative humidity measuring devices and temperature sensors are installed in their housing on the duct before and after the test section. The wind tunnel is equipped with a variable speed fan to achieve variable air flow velocities. Two computers are used; one is connected to the data acquisition system for the air temperature and relative humidity measurements and the other for air velocity measurements. Air velocity is measured using an 8-channel hot wire anemometer. 3. Experimental procedure Control of air flow is established using a variable speed motor controller by adjusting the power source frequency which is fed to the fan motor. Water flow rate is adjusted by a water control valve. Air velocity is measured by hot wire anemometers. Seven velocity sensors are fixed on vertical strips at vertical y-distances 7.42, 28.4, 51.24, 70.0, 88.76, 111.58 and 132.58 mm to scan the air velocity. The vertical strip holding the sensors slides along the horizontal direction of the rectangular cross section to scan the velocity in the horizontal direction. The coordinates of the sensors are shown in Fig. 5. Following this procedure, air velocities are measured at 42 points over the entire cross section. A data acquisition system is connected to a PC and used to collect air velocities. The mean air velocity is estimated based on the 42 measured values. Water flow rate through the test section is measured using a turbine flow meter. Temperature along the air duct is measured using K-type thermocouples and thermistors. Four thermocouples are used, two of them at the inlet and the others at the outlet to measure the dry and wet bulb temperatures. Two thermocouples are fixed before and after the test section to measure the air temperatures across the test section. Two more thermocouples are used to measure the water temperatures in the water tank and at the outlet of the test section. One thermocouple is used to measure the room temperature. Humidity sensors are used to measure the relative humidity before and after the test section. These sensors are also equipped with thermistors to measure temperatures. Two pressure transducers are used to measure the pressure drop across the test section. Another data acquisition system is connected to a laptop to collect the data of humidity sensors, pressure transducers and thermocouples. Table 1 shows the uncertainty in the measured parameters. The thermocouples and data acquisition system are calibrated at ice and boiling points. At the ice point, the error in thermocouple readings is in the range 0.2 to 0.1 C. However, the error at the boiling point is in the range 0.15e0.4 C. The used flow meter is a turbine type (FTB602) made by Omega Engineering, INC. The flow meter is
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Consequently, the data can be represented in terms of Colburn heat transfer group JH (Zhukauskas [16]):
Table 1 Uncertainty of the measured parameters. Temperature Relative humidity Pressure Velocity Water flow rate
5
0.2 to þ0.4 C 2% R.H 0.25% 2% 2%
JH ¼
Nu Re Pr 1n
(8)
Similarity between heat and mass transfer leads to (Zhukauskas [16]):
calibrated using a balance scale and a stop watch. The difference between the reading of the flow meter and that of the measured flow rates is 2% at most.
JD ¼
Sh Re Sc1n
(9)
4. Heat and mass transfer analysis In this section, an attempt is made to analyze the heat and mass transfer between the ceramic tubes and the flowing air. The heat transferred from the air through the membrane tube bank is used to evaporate the water through the membrane surface. Therefore, the heat lost from the air due to water evaporation is given by:
_ a Cp ðTdi Tdo Þ Qw Qa ¼ m
(1)
where Tdi and Tdo are air temperatures at test section inlet and outlet respectively, and Qw is the heat gained by the water flow inside the membrane tube:
_ w Cw ðTiw Tew Þ Qw ¼ m
(2)
where Tiw and Tew are water temperatures at inlet and outlet respectively. The evaporation heat transfer coefficient is estimated from:
he ¼
Qa AðTm Twb Þ
(3)
where A is the tube bank surface area, Twb is the wet bulb temperature which represents the film temperature on the membrane surface and Tm is the mean temperature defined by:
Tm ¼
Tdi þ Tdo 2
(4)
Nusselt number is defined by:
Nu ¼
he d k
(5)
where d is tube diameter and k is the air thermal conductivity. The Reynolds number based on the maximum velocity through the tube bank is calculated from:
Re ¼
rUmax d m
(6)
The general average correlation for the entire tube bundle has been reported by Zhukauskas [16] of the form
0:2 S Nu ¼ F 0:35 T Re0:6 Pr n ðPr=Prs Þ0:25 SL
(7)
This equation is valid for 1000 Re 200,000 and ST and SL are the distances between tube centers normal and parallel to the flow through the tube bank, respectively. F is a correction factor for banks with less than 16 rows and n ¼ 0.36. For the current experiment: F ¼ 0.95 corresponding to 7 rows and (ST/SL) ¼ 1.
Fig. 6. Temperature profiles for 1 h span of air at the inlet and exit from the duct and at the inlet and exit from the test section at fixed air velocity and for various water velocity; (a) vw ¼ 0.01667 m/s, (b) vw ¼ 0.0287 m/s, and (c) test section inlet and outlet temperature at three different water velocities.
6
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Table 2 Dry and wet bulb temperatures at the inlet and outlet of both the test section and the duct where the evaporative cooling system efficiency estimated for the test section. No Duct inlet
Test section inlet
Test section outlet
Duct exit
Test section efficiency %
Td, C Twb, C Td, C Twb, C Td, C Twb, C Td, C Twb, C 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
36.68 37.47 38.5 39.98 39.59 39.85 42.2 43.18 35.06 39.35 40.76 41.19 35.5 39.15 39.81 40.64 35.76 37.3 36.66 40.38 40.58
15 15.6 15.5 17.5 17.5 16.8 18.5 18.7 16 16 18.5 20 16.3 18 17 18.5 16 16.6 14.5 18.2 17.5
35.86 36.79 37.81 38.52 38.64 39.21 41.61 42.49 34.7 38.82 39 40.63 35.21 38.56 38.96 41.06 35.65 36.96 36.64 40.5 40.68
14.71 15.36 15.26 17.02 17.19 16.59 18.31 18.48 15.88 15.82 17.94 19.83 16.2 17.81 16.72 18.63 15.96 16.49 14.49 18.24 17.53
29.99 30.75 31.04 31.58 29.14 32.33 35 36 29.5 33 32.5 32 26.66 32.26 32.08 35.92 31.8 32.69 33.2 35.25 34.39
13.18 14.36 14.77 16.57 15.4 15.01 17.92 17.39 15.24 16.93 15.95 17.65 12.87 16.05 14.44 17.82 13.97 16.58 13.7 18.04 17.44
29.51 30.59 31.67 31.36 28.02 29.56 32.19 33.31 28.83 30.26 29.8 30.03 27.77 29.26 29.54 34.01 31.88 29.85 28.81 30.55 30.1
13 14.3 15 16.5 15 14 17 16.5 15 16 15 17 13.3 15 13.5 17.2 14 15.6 12 16.5 16
27.75 28.18 30.02 32.28 44.29 30.42 28.37 27.03 27.63 25.30 30.86 41.49 44.98 30.36 30.94 22.92 19.55 20.86 15.53 23.58 27.17
where JD is the Colburn mass transfer group and Sh is the Sherwood number:
Sh ¼
hm d Dv
hm ¼ km =ram
(17)
where ram is the average air density. 5. Results and discussion Measurements of the temperature and relative humidity are taken for the ceramic tube test section shown in Fig. 3 at various air and water velocities. Fig. 6(a, b) shows the temperature profiles for two different water flow velocities of 0.0167 m/s (4.71 107 m3/s) in Fig. 6(a) and 0.0287 m/s (8.11 107 m3/s) in Fig. 6(b) as the time goes for 1 h for fixed air velocity of 3.52 m/s (0.138 m3/s). This figure shows that there is minimum heat leakage from the wall of the wind tunnel since there is not much difference between the temperatures at the duct inlet and the test section inlet. The same can be said about the temperature between the test section outlet and the duct outlet as an average. It should be noted that in Fig. 6(b), at the beginning of the experiment, the outlet temperature of the duct is higher than that of the test section by about 5 C. However when the experiment reaches steady state at time between 15 and 35 min the two temperatures become equal. The reason for initial discrepancy could be that the sensor at the outlet of the duct is affected initially by the lab temperature which is higher than that of the outlet test section temperature. Furthermore, on average, there is about 6 C cooling in the air at the exit from the duct. This temperature drop represents a cooling ratio of about 17% compared to inlet air to the duct which based on an average of 35 C inlet air temperature (Fig. 6(a)). On the other hand, there is a 10 C drop in
(10)
where hm is the convective mass transfer coefficient and Dv is the mass diffusivity of water vapor in air. The Schmidt number Sc is:
Sc ¼
m rDv
(11)
According to the similarity, Sh number can be presented by (Zhukauskas [16]):
Sh ¼
0:25 1n 0:2 Sc S Sc F 0:35 T Re0:6 Sc0:36 Pr Scs SL
(12)
The evaporation rate at tube bank surface can be estimated from:
_ ev ¼ Qa =hfg m
(13)
where hfg is the water latent heat of vaporization and the humidity ratio increase can be obtained from:
_ ev =m _a uo ui ¼ m
(14)
The mass transfer coefficient can be estimated from:
km ¼
_ ev m Aðus um Þ
(15)
where us is the humidity ratio at the tube surface and is obtained from the psychometric chart at the measured wet bulb temperature and um is the average humidity ratio:
um ¼ ðue þ ui Þ=2
(16)
According to ASHRAE [17], the convective mass transfer coefficient hm is related to the mass transfer coefficient km by:
Fig. 7. Effect of water flow rate on relative humidity at two different ambient conditions and for various inlet water velocities; (a) vw ¼ 0.01667 m/s and (b) vw ¼ 0.0287 m/s.
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
7
Fig. 8. Temperature profiles for 1 h span of air at the inlet and exit from the duct and at the inlet and exit from the test section at fixed air velocity; (a) for water velocity 0.0236 m/s and (b) for 0.0347 m/s.
Fig. 9. Effect of water velocity on the relative humidity at two different ambient conditions and for fixed air velocity for; (a) vw ¼ 0.0236 m/s and (b) vw ¼ 0.0347 m/s.
the air leaving the duct as seen in Fig. 6(b) which represents a 25% cooling ratio of the inlet temperature (based on about an average 40 C inlet air temperature). Fig. 6(c) shows the inlet and outlet temperature profiles of the test section at 3.61 m/s air velocity and for three water velocities 0.0167 m/s, 0.0229 m/s and 0.0343 m/s. The average temperature difference drop between the inlet and outlet are 9.5 C, 6.8 C and 6.49 C for the different velocities respectively. It should be noted that the numerical values of the average temperature at the inlet and outlet of the test section can be obtained from Table 2, runs number 5, 6 and 8 corresponding to various water velocities respectively. This figure indicates also that as the water velocity increases the temperature drop between the inlet and outlet decreases as seen for run 5 and 6 (Table 2) for almost the same inlet temperature. Fig. 7 shows the relative humidity for fixed air flow velocity and for several water velocities at various inlet air conditions. The inlet relative humidity starts (at time ¼ 0) at about 9.5% at the inlet to the test section and increases to about 13% at the outlet of the test section as shown in Fig. 7(a). It should be noted that as the time goes during the 1 h experiment both the inlet and outlet relative humidity decrease keeping the difference almost constant as seen in Fig. 7(a). However, for the other ambient conditions, the relative humidity starts at about 2% and increases to about 5.4% at the exit of the test section as seen in Fig. 7(b). This observation confirms the decrease in the air temperature seen in Fig. 6(a, b). Fig. 8(a, b) shows the temperature profiles for suction air velocity 4.26 m/s (0.167 m3/s) and for two water velocities 0.0236 m/s (6.67 107 m3/s) in Fig. 8(a) and 0.0347 m/s (9.81 10 m3/s) in Fig. 8(b). Fig. 8(a) shows that the
ambient inlet duct temperature is almost constant at an average value of 40.38 C. This figure also shows that the average duct outlet temperature is 30.55 C with a drop in air temperature of 9.83 C which represents a 24.3% cooling ratio of the ambient temperature. It is worth mentioning that the duct outlet temperature is lower than the test section outlet temperature, which could be attributed to some small droplets of water that evaporate in the passage between the test section outlet and the duct outlet. A 25.8% decrease in the duct air outlet temperature is noticed in Fig. 8(b) as the water velocity increased to 0.0347 m/s (9.81 107 m3/s). The corresponding relative humidity is shown in Fig. 9(a, b). The average relative humidity increases from about 0.2% to 2.3% as seen in Fig. 9(a and b). Water temperature at the inlet and outlet of the test section is shown in Fig. 10(a) and (b) for two different water velocities and at fixed air speed of 3.52 m/s (0.138 m3/s). As seen in Fig. 10(a) at water velocity 0.01667 m/s (4.71 107 m3/s), the maximum temperature difference between the inlet and outlet is about 2 C. However, no remarkable difference is observed as the water velocity increased to 0.0249 m/s (7.04 107 m3/s) as seen in Fig. 10(b). It should be noted that data in Fig. 10(a) are corresponded to the data given in Figs. 6(a) and 7(a) for temperature and relative humidity respectively. Therefore, Fig. 10 shows that part of the absorbed heat from the air is used as a latent heat for evaporating the water as discussed earlier in the heat and mass transfer analysis section. It should be mentioned that low water flow rates are used such that the water can go through the pores in a vapor form to avoid droplets or film of water at the outside surface of the tubes. On the other hand, it is noticed that the evaporation process on the
8
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
Fig. 10. Effect of water velocity on the water temperature before and after the test section at fixed air velocity; (a) vw ¼ 0.01667 m/s and (b) vw ¼ 0.0249 m/s.
psychometric chart does not follow isenthalpic lines. This means that not all the dissipated energy from the air is transferred as latent heat to the evaporating water but part of it is used as a sensible heat. Samples of the psychometric chart processes are shown in Fig. 11(a, b, c). These plots are based on dry and wet bulb temperatures at the test section inlet and outlet. Fig. 11(a, b, c) corresponds to some runs shown in Table 2. It is clear from these figures that the process curves do not exactly follow the isenthalpic lines; therefore some significant non-evaporative heat transfer is present. Fig. 12 summarizes the relation between the water flow rates through the ceramic pipes for different air velocities and the relative humidity before (BTS) and after the test section (ATS). This figure shows that the relative humidity decreases as the water flow rate increases at higher values of air velocities. This figure also summarizes the rate of increase of the relative humidity RH before (BTS) and after the test section (ATS) as explained earlier in Fig. 7(a,b) and 9 (a,b) for some water and air velocities. Table 2 shows the performance of the experimental data at the test section using the evaporative cooling system efficiency (Tdin Tdout)/(Tdin Twin). It should be noted that the wet bulb temperature at the duct inlet and outlet are measured using sensors covered with wet wick. Wet bulb temperatures at test section inlet and outlet are obtained from the psychometric chart (EES software) assuming that the process from the duct inlet to test section inlet has the same humidity ratio, and that the process from the test section outlet to duct outlet has a constant humidity ratio. Table 3 shows the air pressure losses for different air and water velocities along the test section. As seen in this Table as the air velocity increases the pressure loss increases up to a maximum value of 44 Pa.
Fig. 11. Psychometric chart for some selected processes; numbers in the figure referred to the experimental runs shown in Table 2; (a) runs 1, 2 and 4 (b) runs 5, 6 and 8 and (c) runs 18, 19 and 20.
Comparison between the staggered tube bank heat transfer represented by the Colburn heat transfer group JH Eq. (8) and the current experimental data is shown in Fig. 13. The present experimental data show good agreement with Eq. (8) since they lie within 20%. Furthermore, comparison between the predicted Colburn mass transfer group JD Eq. (9) and the current experimental data is shown in Fig. 14. Again our experimental data is within 20% of correlation Eq. (9). 6. Conclusions Results show that using the modified ceramic tubes enhances the heat transfer between the air and water vapor. This enhancement is reflected by a maximum and minimum temperature drop of 11 C and 4 C of the dry air respectively. It is noticed that some distance should be left between the test section outlet and the duct outlet to give a chance for any droplet of water already in the air to evaporate before exiting the duct for further use. As expected, the relative humidity downstream of the test section is found to increase by about 3.5% compared to ambient conditions. Heat and mass transfer analysis for the tested ceramic tubes arrangement compared within 20% with the published (8, 9) correlations for
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
9
Fig. 14. Comparison of the current experimental data with the Colburn mass transfer group (9) for staggered tube bank.
Fig. 12. The effect of water flow rate on the relative humidity before and after the test section for various air velocities.
Table 3 Air pressure losses at various values of air and water at the test section. No.
Vw m/s
Va m/s
DP Pa
1 2 3 4 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
0.0167 0.0226 0.0287 0.0344 0.0229 0.0287 0.0343 0.0169 0.0231 0.0286 0.0340 0.0170 0.0225 0.0280 0.0343 0.0330 0.0406 0.0163 0.0236 0.0347
3.5188 3.5188 3.5188 3.5188 3.6145 3.6145 3.6145 4.0334 4.0334 4.0334 4.0334 3.8031 3.8031 3.8031 3.8031 4.2575 4.2575 4.2575 4.2575 4.2575
7.79 22.15 10.33 13.35 10.13 10.60 10.96 22.06 17.33 23.99 28.21 8.36 12.36 18.60 16.80 17.85 25.57 45.73 36.47 44.00
Fig. 13. Comparison of the current experimental data with the Colburn heat transfer group JH (8) for staggered tube bank.
the staggered tubes arrangement. It is noticed that the evaporation process does not follow the isenthalpic lines; therefore, heat is transferred from the air as latent and sensible heats. The experimental data show that ceramic tube membrane technology could be implemented for gas turbine inlet air cooling without concerns of compressor blades erosion due to water droplets associated with water spray technology (also known as fogging technology). However, more research needs to be performed to improve the efficiency of the proposed technology. Acknowledgments The authors would like to acknowledge the support of National Plan for Science, Technology, & Innovation (NPST) at King Saud University under project number ENE220-02-08. This support is highly appreciated and acknowledged. References [1] N. Ravi Kumar, K. Rama Krishna, A.V.S. Rama Raju, Improved gas turbine efficiency using spray cooler and alternative regenerator configuration, in: ICRAMME 05, Proceedings of the International Conference on Recent Advances in Mechanical &Materials Engineering, 30e31 May 2005. Kuala Lumpur, Malaysia, Paper No. 10. [2] S. Jolly, J. Nitzken, D. Shepherd, Direct Spray System for Inlet Air Cooling W 501 B5, Presented at the Power-Gen International, Dallas, Texas, December 9e 11, 1998. [3] D.C. Erickson, I. Kyung, G. Anand, E.E. Makar, Aqua absorption turbine inlet cooler, in: Proceedings of ASME International Mechanical Engineering Congress & Exposition Washington, D.C, November 16e21, 2003. [4] D.V. Punwani, R. Pasteris, GT inlet-air cooling boosts output on warm days to increase revenue, Comb. Cycle J. Fourth Quarter (2003) 1e3. [5] M.M. Alhazmy, Y.S.H. Najjar, Augmentation of gas turbine performance using air coolers, Appl. Therm. Eng. 24 (2004) 415e429. [6] M. Ameri, S.H. Hejazi, The study of capacity enhancement of the Chabahar gas turbine installation using an absorption chiller, Appl. Therm. Eng. 24 (2004) 59e68. [7] S. Loeb, Membrane evaporative cooling to 30 C or less, “Membrane evaporative cooling of contained water, Ann. N. Y. Acad. Sci. 984 (March 2003) 515e 527. Advanced Membrane Technology. [8] D.W. Johnson, C. Yavuzturk, J. Pruis, Analysis of heat and mass transfer phenomena in hollow fiber membranes used for evaporative cooling, J. Membr. Sci. 227 (1e2) (2003) 159e171. [9] D. Bessarabov, Z. Twardowski, New opportunities for osmotic membrane distillation, Membr. Technol. (7) (2006) 7e11. [10] E. Drioli, V. Calabrd, Y. Wu, Microporous membranes in membrane distillation, Pure Appl. Chem. 58 (12) (1986) 1657e1662. [11] F. Tillberg, ZLD-Systems An Overview, Department of Energy Technology, Royal Institute of Technology, KTH, Stockholm, 2004. [12] H.J. Zwijnenberg, G.H. Koops, M. Wessling, Solar driven membrane evaporation for desalination processes, J. Membr. Sci. 250 (2005) 235e246. [13] L. Zhang, Performance deteriorations from flow maldistribuion in air-to-air heat exchangers: a parallel-plates membrane core case, Numer. Heat Transfer Part A 56 (2009) 746e763.
10
O. Zeitoun et al. / International Journal of Thermal Sciences 80 (2014) 1e10
[14] J. Lu, W. Lu, A numerical simulation for mass transfer through the porous membrane of parallel straight channels, Int. J. Heat Mass Transfer 53 (2010) 2404e2413. [15] A. Sommers, Q. Wang, X. Han, C. T’Joen, Y. Park, A. Jacobi, Ceramics and ceramic matrix composites for heat exchangers in advanced thermal systemsA review, Appl. Therm. Eng. 30 (2010) 1277e1291.
[16] A. Zhukauskas, Heat transfer from tubes in cross flow, in: J.P. Hartnett, T.F. Irvine Jr. (Eds.), Advances in Heat Transfer, vol. 8Academic Press, New York, 1972. [17] ASHRAE, Chapter 6 Mass Transfer, ASHRAE Handbook e Fundamental, 2009.