- Email: [email protected]
Modified pulse operation of thermoelectric coolers for building cooling applications
Energy Conversion and Management 140 (2017) 145–156
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Modified pulse operation of thermoelectric coolers for building cooling applications S. Manikandan a,⇑, S.C. Kaushik a, Ronggui Yang b a b
Centre for Energy Studies, Indian Institute of Technology Delhi, India Department of Mechanical Engineering, University of Colorado at Boulder, CO, USA
a r t i c l e
i n f o
Article history: Received 11 January 2017 Received in revised form 28 February 2017 Accepted 1 March 2017
Keywords: Thermoelectric cooler Pulse operation Building cooling
a b s t r a c t This paper presents a modified pulse operation of the thermoelectric cooler (TEC) for building space cooling application. In the normal pulse operation of a thermoelectric cooler, a current pulse is given as the input, but in the modified pulse operation, along with the pulse current, the hot side heat transfer coefficient is also pulsed. A numerical model for the thermoelectric cooling system is developed assuming one-dimensional unsteady state heat transfer having convective heat transfer boundary condition at its hot side and constant heat flux boundary condition at its cold side. Then the thermoelectric cooling system is studied with variable pulse current ratio, cooling load, variable pulse width and with dissimilar pulse shapes. The result shows that, for a typical operating condition with the modified pulse operation, the thermoelectric cooling system can provide an average cooling power of 600 W with the COP of 1.01, which are 23.3% and 2.12% higher than the normal mode of operation (i.e. without current pulse) respectively. This study also found that the rectangular-shaped pulse can provide higher average cooling power and COP when compared with the ramp and exponential pulse. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction The use of thermoelectric coolers for building space cooling can be an alternative option to the conventional vapour compression refrigeration (VCR) systems, since it does not use any refrigerants, it is compact and provides noiseless operation. However, its coefficient of performance (COP) is lower when compared with the conventional VCR systems due to the low figure of merit (Z) of thermoelectric materials [1]. The thermoelectric cooling systems have been compared with the conventional vapour compression cooling system, vapour absorption cooling system and Stirling cycle cooling systems and it was found that the COP of the thermoelectric cooler system is lower when compared with the other above mentioned cooling systems [2–4]. These studies also suggested that, if the figure of merit of the thermoelectric materials are improved from the present commercially-available bismuth telluride alloy based thermoelectric coolers, which has an average dimensionless figure of merit (ZT) of 0.7 to higher than 3. Then the COP and the cooling power of thermoelectric coolers can be improved significantly, with a great potential to replace the con-
⇑ Corresponding author. E-mail address: [email protected] (S. Manikandan). http://dx.doi.org/10.1016/j.enconman.2017.03.003 0196-8904/Ó 2017 Elsevier Ltd. All rights reserved.
ventional air-conditioning systems for building space cooling applications. The design of thermoelectric coolers using number of transfer units for electronic cooling has been proposed by Cai et al. [5]. Wang et al. [6] have studied the cooling of light emitting diode using a thermoelectric cooler package, and found that the thermoelectric device is more sensitive to the operating current. Attar and Lee [7] have theoretically studied and experimentally verified the thermoelectric cooler for automotive cooling application with the cooling power of 400 W and COP of 1.27. Maneewan et al. [8] have experimentally studied a compact thermoelectric air conditioner with the operating temperature of 301 K, cooling power of 29 W and COP of 0.34, and also estimated the payback time of this system to be 0.75 years. Cosnier et al. [9] have numerically and experimentally studied the thermoelectric cooling and heating systems and found that the COP of the thermoelectric cooler may vary from 1 to 0.3 when the temperature difference between its hot and cold side varies from 15 K to 30 K. Gillott et al. [10] have experimentally studied the thermoelectric air-conditioner with the average cold side temperature of 295 K, cooling load of 220 W and COP of 0.46. To improve the COP of thermoelectric cooling systems for building cooling applications, researchers have integrated various techniques such as thermal energy storage and evaporative cooling with the thermoelectric air-conditioning system. Tipsaenporm et al. [11] have used the direct evaporative cooler to supply cold
146
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Nomenclature h m A C D I K L M N P Q R T U
heat transfer coefficient (W m2 K1) mass (kg) area (m2) specific heat (J kg1 K1) density (kg m3) current (A) thermal conductance (W K1) length number of infinitesimal sections of the thermoelement number of thermoelectric cooler modules electrical power (W) heat (W) electrical resistance (X) temperature (K) overall heat transfer coefficient (W m2 K1)
Greek letters a seebeck coefficient (V K1) k thermal conductivity (W m1 K1) q electrical resistivity (X m) r electrical conductivity (S m1) D difference air to the hot side of the thermoelectric cooler to assist the heat removal. This technique has improved the cooling power of the thermoelectric cooler by 40.6% and the COP by 20.9%. However, the COP of this system was still low as 0.52 with the cooling load of 74.5 W. Tan and Zhao [12] have studied the integrated thermal energy storage-assisted thermoelectric cooling system where the phase change material is used to store the cold energy during the night and then that coldness is used in the day time to assist the heat rejection from the hot side of the thermoelectric cooler with average cold side temperature of 296 K. This technique have improved the COP of the thermoelectric cooler from 0.5 to 0.78 (56%). Even though, these techniques improved the COP of the thermoelectric cooler, the integration of evaporative cooler/ thermal energy storage makes the system bulky and adds to the operation complexity. The solar operation of the thermoelectric cooler for building cooling application with the cooling load of 589 W for a room sized 2.8 2.7 2.6 m3 has been studied by Irshad et al. [13,14]. This system has a COP of 0.68 and the cooling load of 498 W with the average cold space temperature of 298 K. The pulse operation of the thermoelectric cooler can provide cold side temperature lower than the cold side temperature at steady state conditions. When the pulse current is applied to the thermoelectric cooler, the cold side temperature drops immediately because of Peltier cooling and then it increases to a peak value and gets peak overshoot because of accumulation of Joule heat in the thermoelement and Fourier heat conduction through the thermoelements from hot side to the cold side. Then the peak overshoot in cold side temperature decays to its steady state value as shown in Fig. 2(b). The cold side temperature has such profile, because the reaction time of Peltier cooling in the junction is much faster than the Fourier heat conduction. Snyder et al. [15] have numerically and experimentally studied the pulse thermoelectric cooling and showed that this technique can provide low cold side temperature which is comparable to the cold side temperature of a two-stage thermoelectric cooler. Yang et al. [16] have derived the expressions for the time to reach the minimum temperature and the holding time with the pulse operation of thermoelectric cooler. This study also analysed the influence of pulse shape and the thermoelement with variable cross section area on the temperature profile of the thermoelectric cooler. Ma et al. [17] have studied
Subscripts a environment b base plate c cold side of TEC dt infinitesimal time dx infinitesimal length gen generation h hot side of TEC i node number in input m mean/average n n type material o reference out output p p type material t time C ceramic Superscripts n incremental time
the continuous pulse operation of the thermoelectric cooling system and found that the periodic super cooling can be obtained, if the TEC is operated in the continuous current pulse mode. Shen et al. [18] and several others [19–22] have theoretically studied and experimentally demonstrated the pulse operation of the thermoelectric coolers. All these studies on pulse operation of thermoelectric coolers are envisioned to make use of the lower cold side temperature for thermal management of electronics/lasers. 1.1. Proposed methodology The studies on pulse operation of thermoelectric coolers proved that it can provide excess cooling for the period of the pulse width. Therefore, if the excess cooling (which is at lower temperature) is utilized at relatively higher temperature (288 K) for the building cooling applications, the average cooling power can be improved which would result in the improved COP. It was found that the thermoelectric cooling systems for building cooling applications have relatively lower COP. It was also identified that, if the cooling power of the thermoelectric cooler increases, the COP will decrease, but, for building cooling applications, the COP and the cooling power are equally important. It can be argued that the cooling power of the thermoelectric cooling system can be increased by adding more number of thermoelectric modules. However, it should be noted that, the effect of adding more number of modules to improve the COP of the system is uncertain and it also increase the capital cost of the system. Therefore, in this paper a new operating technique of the thermoelectric cooler is proposed which can improve the COP and cooling power of the thermoelectric cooler simultaneously without adding more thermoelectric modules. This is the novelty of the present research. 2. Thermodynamic modelling for pulse operation of thermoelectric cooler Fig. 1 shows the schematic of the thermoelectric cooler system for the building space cooling application. The thermoelectric properties and the dimensions of the thermoelectric device are given in Table 1. Here thermoelectric properties of commercially available bismuth telluride (Bi2Te3) has been used in this study.
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
147
Fig. 1. Schematic representation of the thermoelectric cooler for building space cooling application.
Fig. 2. Input variables and cold side temperature profile of normal pulse operation (a and b) and modified pulse operation (c and d) of TEC.
The following assumptions have been made for the modelling and analysis in this work: (1) One dimensional unsteady state heat transfer along the length of the thermocouple has been followed.
(2) Averaged thermoelectric material properties have been used for this study. (3) No convective and radiative heat transfer from the sides of the thermocouples to the surroundings. (4) The thermal and electrical contact resistances are ignored.
148
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Table 1 Properties of the thermoelectric cooler [23]. Sl. no.
Material parameter 1
1 2 3 4 5 6 7 8 9 10 11 12 13 14
1
Cp, Cn (J kg K ) CC,c, CC,h (J kg1 K1) Cb,c, Cb,h (J kg1 K1) mC,c, mC,h (kg) mb,c, mb,h (kg) Dp, Dn (kg m3) Ta (K) kp + kn (W m1 K1) qp + qn (X m) ap an (V K1) Ap, An (m2) Lp, Ln (m) N Number of thermocouples per module
Value 200 419 400, 850 2.95 105 8.64 105 10922.08 300 3.363 2.10 105 420 106 12.25 106 4.25 103 24 127
temperature at an incremental time (T nþ1 ) is a function of its temo perature at its previous time step (T no ), temperature of its nearest
2.1. Finite difference modelling of the thermoelectric cooler The total number of thermoelectric cooler modules used in this study is 24 and each thermoelectric cooler has 127 thermocouples. To simplify the numerical modelling, equivalent stage method has been assumed [24]. The one-dimensional unsteady state heat transfer model of the thermoelectric cooler has been developed by applying the energy balance at its hot side, cold side and in a small cross section along the length of the thermoelement. Then the finite difference approximation is applied to the energy balance equations to obtain the temperature profile of the thermoelectric cooler. The energy balance at the cold side of the flat thermoelectric cooler can be given as by Cheng et al. [25]
ðmb;c C b;c þ mC;c C C;c Þ
dT c ¼ Q c þ Q k IaT c dt
ð1Þ
where mb,c and mC,c are the mass of base plate and ceramic layer at cold side respectively and Cb,c and CC,c are the specific heat of base plate and ceramic layer at the cold side respectively. The similar nomenclature have also been followed at the hot side of the thermoelectric cooler. The Fourier heat conduction term Qk in the above equation can be given by Eq. (2), Qc is the cooling load and a can be defined as the combined Seebeck coefficient (a = ap an).
Q k ¼ kA
dT dx
ð2Þ
The energy balance at the hot side of the thermoelectric cooler can be given as
ðmb;h C b;h þ mC;h C C;h Þ hAs ðT h T a Þ ¼ h
dT h ¼ IaT h Q k hAs ðT h T a Þ dt
As AðT h T a Þ ¼ UAðT h T a Þ A
ð3Þ ð4Þ
where As is the surface area of the hot side heat sink, A is the cross section area of thermoelectric couples (A = 24 thermoelectric cooler modules 127 thermocouples Area of single thermocouple), and U is the overall heat transfer coefficient at the hot side of the thermoelectric cooler. The energy balance in a small section of the thermoelement (n type or p type) using the 1-D unsteady state heat transfer modelling can be given as 2
d T 2
dx
þ
I2 q kA
2
¼
DC p dT dt k
and node numbers from i = 0 to i = M + 1, where, Δx can be defined as the ratio of length of the thermocouple (L) to the number of sections (M). Then, by applying the finite difference approximation in the Eqs. (1)–(4) by following simple implicit method, we obtain two equations for the cold and hot side boundary temperatures and M 1 number of equations for the nodal temperatures of the thermocouples, accounting to M + 1 number of finite difference equations. Simple implicit finite difference method has been used to develop the finite difference equations, since it is always stable and more accurate when compared with the simple explicit method. Further details about the finite difference technique, simple implicit method, and the techniques to solve the finite difference equations can be referred from Ozisik [26]. The cold side temperature at boundary node i = 0 can be calculated using Eq. (5). It can be seen from the Eq. (5) that the cold side
ð5Þ
The thermoelements in the thermoelectric cooler has been divided into M number of equal parts with equivalent length (Δx)
node at an incremental time (T nþ1 1 ) and also the boundary condition (Qc).The terms (mb;c C b;c þ mC;c C C;c ) in the Eq. (1) and (mb;h C b;h þ mC;h C C;h ) in Eq. (3) have been expressed as M c C c and Mh C h respectively in subsequent equations wherever applicable.
M c C c 2Dx 1 Ia2Dx 1 2 1 þ þ þ kADt kA Dx2 Dx2 Dx2 F Dt 2 M c C c 2Dx 1 1 þ T no þ þ T nþ1 1 kADt F Dt D x2 Dx2 Q c 2 Dx 1 þb ¼0 ð6Þ þ 2 kA Dx
T nþ1 o
It should be noted that the subscripts (i) of the temperature (T) in Eqs. (6)–(8) corresponds to the node number and superscripts (n, n + 1,. . .) corresponds to the time and incremental time respectively. The hot side temperature of the thermoelectric cooler at an incremental time (i.e. node number i = M) can be calculated using Eq. (6). It can be seen from the below equation that, the hot side temperature at an incremental time (T nþ1 M ) is a function of its temperature at previous time step (T nM ), temperature of its nearest node at an incremental time (T nþ1 M1 ), the hot side heat transfer coefficient and the environmental temperature. M h C h 2Dx 1 Ia2Dx 1 2 1 þ T nþ1 þ M kADt kA Dx2 Dx2 Dx2 F Dt hAs 2Dx 1 2 M h C h 2Dx 1 1 nþ1 n þ T þ T þ þ M1 M kA kADt F Dt Dx2 Dx2 Dx2 T a UA2Dx 1 þ þb ¼0 ð7Þ kA Dx2 For all the nodes other than the boundary nodes, the temperature distribution in the thermoelement (i.e., for i = 1 to M 1) can be given as
2 1 1 1 Tn nþ1 nþ1 þ T þ i þb¼0 þ þ T i1 iþ1 Dx2 F Dt Dx2 Dx2 F Dt
T nþ1 i
ð8Þ
The constants (b and F) mentioned in Eqs. (6)(8) can be given as follows:
b¼
I2 q kA
2
and F ¼
k DC p
The power input to the thermoelectric cooler can be calculated as
Pin ¼ aIðT h T c Þ þ I2
qL A
ð9Þ
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Then the COP of the thermoelectric cooling system considering the input power required by the fans at its hot and cold sides can be given as
COP ¼
Q c;av erage Pin þ P fan
ð10Þ
where Qc,average, Pin and Pfan can be defined as
Z tþDtpulse þtsteady 1 Q c ðtÞ dt; Dt pulse þ t steady t Z tþDtpulse þtsteady 1 Pin ðtÞ dt and Pin ¼ Dtpulse þ tsteady t Z tþDtpulse þtsteady 1 P fan ðtÞ dt Pfan ¼ Dt pulse þ t steady t
Q c;av erage ¼
where tsteady is the time taken to reach the steady state after the application of current pulse and Dtpulse is the duration of pulse width. The power input required by a fan has been assumed to be 14 W and a total 8 fans have been used in this study (4 each at the hot and cold side of the thermoelectric cooler). Here 4 fans are operated during the normal operation (without any pulse) to facilitate the heat rejection from the hot side of the thermoelectric cooler and to supply cold air to the cooling space respectively, and during the pulse operation (i.e. only for the period of Dtpulse), additional 4 fans (total of 8) have been operated to increase the cooling power and hot side heat transfer coefficient (refer Fig. 2(c)). 2.2. Modified pulse operation of thermoelectric cooler The difference between the normal pulse operation and the modified pulse operation of the thermoelectric cooler is shown in Fig. 2. In the normal pulse operation, a pulse current (Ipulse) large than the steady state current will be given for a period of time (Δtpulse) in addition to a steady state current (Isteady) after the thermoelectric cooler reached its steady state. When the current pulse is applied, the temperature of the cold side suddenly drops to a temperature lower than the temperature of the single stage thermoelectric cooler at steady state condition. It should be noted that, during the normal pulse operation, the hot side temperature of the thermoelectric cooler for the pulse period is higher than hot side temperature at steady state condition. The increase in hot side temperature may influence the cold side temperature of the thermoelectric cooler because of the excess Joule heat accumulated in the thermoelement for the pulse period and Fourier heat conduction through the thermoelement and this causes a higher peak overshoot in the cold side temperature, as shown in Fig. 2(b). Therefore, in the modified pulse operation, to reduce the peak overshoot in the cold side temperature, the hot side heat transfer coefficient should also be increased for a period of Δtpulse as shown in Fig. 2(c), which minimize the temperature rise in hot side of the thermoelectric cooler (during the pulse period) and therefore, reducing the peak overshoot in the cold side temperature. In building space cooling applications with thermoelectric coolers, the cold side temperature can be 288 K [11], so that the supply air to the cold space will be 296–298 K as per the thermal comfort level [27]. Therefore, in the modified pulse operation, the cooling power will be increased with the applied pulse current, so that the larger temperature drop in the pulse mode can be utilized to cater more cooling load, which increases the cold side temperature up to the required level for space cooling application. In practice, the pulse hot side heat transfer coefficient and the pulse cooling power can be achieved by increasing the mass flow rate of the fluid flowing through the hot and cold side heat exchangers of the thermoelectric cooler.
149
3. Results and discussion The cooling load requirement for a room sized 2.8 2.7 2.6 m3 with the thermoelectric cooling system has been evaluated to be 589 W [13]. To select the initial values for the operating current, the hot side heat transfer coefficient and the base cooling load, the thermoelectric cooling system has been analysed at steady state without any pulse current. The variation in steady state cold side temperature and COP of the thermoelectric cooler with the hot side heat transfer coefficient for a constant cooling load has been presented in Fig. 3. It shows that, with the increase in hot side heat transfer coefficient, the cold side temperature of the thermoelectric cooler decreases, this is because of the decrease in the hot side temperature of TEC. Thus, increasing the hot side heat transfer coefficient effectively reduces the temperature difference between the hot and cold side of the thermoelectric cooler (Th-Tc). After a certain value of hsteady, the rate of change of Th-Tc decreases, thus making the COP increase at smaller rate. Therefore, hsteady has been chosen as 200 W m2 K1, and this value agrees with Wang et al. [28]. The hot side heat transfer coefficient during the pulse mode (hpulse) is then assumed to be 2 times of hsteady. Fig. 4 shows the variation on cooling power and COP of a single thermoelectric cooler with different operating current (without any pulse) with hot side heat transfer coefficient fixed at 200 W m2 K1. With the increase in current, the COP of the thermoelectric cooler increases first to reach its maximum and then decreases as expected [29]. In the normal pulse operation, the thermoelectric cooler are operated at the current for maximum temperature difference between its hot and cold junction. In building cooling applications it is important to maintain the cold space temperature. Therefore, the operating current should be lower than the optimum current for maximum temperature difference. Hence, in this case, when the temperature difference between the hot and cold side of the thermoelectric cooler is 12 K, Isteady has been chosen as 4 A, with which the COP is maximum. With the operating current of 4 A, and the number of thermoelectric cooler modules of 24, the steady state cooling load can be 486 W.
3.1. Influence of pulse cooling load The variation in cold side temperature of thermoelectric cooler under different pulse cooling load with constant pulse width of 5 s is presented in Fig. 5. The pulse cooling load is the cooling load of the thermoelectric cooler during the current pulse period, which has been selected to meet the average cooling load requirement of 600 W. Thus, the pulse cooling load has been selected as 1850 W. Three points can be noted from Fig. 5: (i) as the Qc,pulse decreases, the cold side temperature during the period of current pulse reduces and vice versa if the pulse cooling load increases, (ii) as the pulse cooling load increases, the peak overshoot in the cold side temperature increases so that the average cold side temperature also increases, (iii) as the pulse cooling load increases, the time taken to reach the steady state increases. The cold side temperature of the thermoelectric cooler decreases with the decrease in pulse cooling load, because at higher currents thermoelectric cooler can cater more cooling load for a fixed cold side temperature. However, in this case, the operating current increases (pulse current) and the cooling load is low, which leads to the decrease in the cold side temperature. This statement can be analytically verified using Eq. (6). Moreover, the increase in the cooling load increases the temperature difference between the hot and cold side of the thermoelectric cooler, which results in increased Fourier heat conduction through the ther-
150
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Fig. 3. Variation of cold side temperature and COP of TEC with hot side heat transfer coefficient at steady state.
Fig. 4. Variation on cooling power and COP of TEC as a function of steady state current.
moelement and thus, increasing the peak overshoot in the cold side temperature at higher pulse cooling loads. If the peak overshoot is higher with constant heat transfer coefficient at the hot side, it takes longer time to reach back to the steady state. The average cooling power has been calculated by integrating the cooling power including the pulse cooling power for the period of 60 s (i.e. (t + Dtpulse + tsteady) (t) = 60 s). The time frame of 60 s have been used because, the cold side temperature reaches the steady state before 60 s. It can be argued that reducing the time frame may increase the average cooling load and COP, but by doing so, the average cold side temperature will increase than the required cold side temperature and also the cold side temperature may not reach the steady state temperature (after applying the current pulse). If the steady state temperature is not reached before applying second (or) consecutive current pulse, the average cold side temperature of the TEC will increase further and it is not
desirable. The average COP has been calculated as the ratio of cooling power (integrated over the period of 60 s) to the total input power integrated over the same period of time as mentioned in Eq. (10). Fig. 6 shows that the average cooling power and COP of the thermoelectric cooler increases with increase in the pulse cooling power. With the increase in pulse cooling power, the average cooling power and average cold side temperature increases, which results in the decrease in the temperature difference between the hot and cold sides and the increase in the COP. 3.2. Influence of pulse current ratio The influence of pulse current on the thermoelectric cooler has been studied with constant average cooling power and with the average cold side temperature, in the former the average cold side
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
151
Fig. 5. Cold side temperature of TEC with modified pulse operation (hsteady = 200 W m2 K1, hpulse = 400 W m2 K1, Isteady = 4 A, Ipulse = 8 A, Qc,steady = 486 W, and Δtpulse = 5 s).
temperature varies where in the later, the average cooling power varies. The pulse current ratio is defined as the ratio of the pulse current to the steady-state current. 3.2.1. Influence of pulse current ratio with constant average cooling power The cold side temperature of the thermoelectric cooler with different pulse current ratio for constant average cooling power of 600 W is presented in Fig. 7(a). It shows that the average cold side temperature decreases with the increase in pulse current ratio, which is because, at higher currents thermoelectric cooler can cater more cooling load. However, in this case, the operating current increases (pulse current) and the cooling load is constant, which leads to the decrease in the cold side temperature. The COP and average cold side temperature of the thermoelectric cooler for different pulse ratio with constant average cooling
load of 600 W are presented in Fig. 8. It can be seen that the COP decreases with the increase in pulse current. This can be explained with two reasons: (i) as the pulse current increases, the Joule heating in the thermoelements increases and thus, reduce the COP (ii) if the pulse current decreases, the average cold side temperature increases, which results in decreased temperature difference between the hot and cold side of the thermoelectric cooler and therefore, the COP of the thermoelectric cooler increases. Since the COP decreases linearly with the increase in pulse current ratio, the optimum pulse ratio can be selected based on the average cold side temperature to match the temperature requirement for building space cooling application. If the pulse current ratio is chosen below 2, the COP will increase, but the cold side temperature of TEC also increases and it in turn increase the cold space temperature. The optimum pulse ratio has been selected as 2. The COP of the thermoelectric cooler with the pulse current ratio of 2 is 1.01
Fig. 6. Average cooling power and COP as a function of pulse cooling load.
152
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Fig. 7. Cold side temperature vs. the pulse current ratio (a) with constant average cooling power (hsteady = 200 W m2 K1, hpulse = 400 W m2 K1, Isteady = 4 A, Qc,steady = 486 W, Qc,pulse = 1850 W, and Δtpulse = 5 s), (b) with constant average cold side temperature (hsteady = 200 W m2 K1, hpulse = 400 W m2 K1, Isteady = 4 A, Qc,steady = 486 W, and Δtpulse = 5 s).
Fig. 8. Variation in COP and the average cold side temperature of the thermoelectric cooler with pulse current ratio.
which is 2.12% higher than the COP of thermoelectric cooler under normal operating condition.
3.2.2. Influence of pulse current ratio with constant average cold side temperature The influence of pulse current ratio has also been studied with constant average cold side temperature of 288.6 K (i.e. the temperature difference between the average cold side temperature and the steady state cold side temperature is zero). It can be seen from Fig. 7(b) that with the increase in pulse current ratio the pulse cooling load should increase to keep the average cold side temperature constant. However, the cold side temperature oscillation increases which is not desirable for building space cooling application since it adversely affect the thermal comfort of the occupants. Therefore, it is not advised to operate the thermoelectric cooler with higher pulse current. Moreover, at the higher pulse current, the COP of the thermoelectric cooler decreases, even if the averaged cooling power increases as shown in Fig. 9.
3.3. Influence of pulse width The variation in cold side temperature of the thermoelectric cooler with different pulse width is presented in Fig.10. It can be seen that the peak overshoot in the cold side temperature of thermoelectric cooler increases if the pulse width increases. The hot side temperature of the thermoelectric cooler increases for the period of pulse width results in increased accumulation of Joule heat in the thermoelement and increases Fourier heat conduction through the thermoelement, thus increasing the peak overshoot. It should be noted that the minimum cold side temperature is constant, due to the constant current pulse ratio. It can be seen that the cold side temperature reaches its steady state before 60 s for all cases, but the increase in pulse width causes longer time to reach the steady state and the longer decay time is not desirable for building space cooling applications. The variation of the average cooling power, the COP and the average cold side temperature of the thermoelectric cooler with the pulse width is shown in Fig. 11. As the pulse width increases,
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
153
Fig. 9. Variation of average cooling power and COP of TEC with current pulse ratio and pulse cooling power.
Fig. 10. Cold side temperature of TEC for different pulse width (hsteady = 200 W m2 K1, hpulse = 400 W m2 K1, Isteady = 4 A, Ipulse = 8 A, Qc,steady = 486 W, and Qc,pulse = 1850 W).
the average cooling power and the average cold side temperature increase, which results in increased COP. Since the average cooling power, COP and average cold side temperature increase with the increase in pulse width, the optimum pulse width can be selected based on the acceptable average cold side temperature. The optimum pulse width has been found as 5 s even though, increasing the pulse width increases the COP and average cooling power. It should be noted that, with the increase in the pulse width, the average cold side temperature increases because of the increase in peak overshoot and it is not desirable. For a typical operating condition with pulse width of 5 s, the increase in average cooling power and COP are 23.3% and 2.12% respectively when compared with the normal mode of operation. 3.4. Influence of pulse shape The pulse shape effect on the modified pulse operation of thermoelectric cooler has also been studied. The influence of ramp up,
ramp down, exponential rise, exponential decay and square (rectangular) pulse on the cooling power and COP of the thermoelectric cooler have been studied. The schematic of ramp up, exponential rise and square pulse is presented in Fig. 12. The ramp down and exponential decay pulse can be the mirror image of the ramp up and exponential rise pulse with respect to the time axis. The pulse width for all the pulses has been maintained as 5 s and the square pulse has been used as the base case for the comparison. It should be noted that the cooling power and the heat transfer coefficient at the hot side follows the shape of the current pulse in the respective cases. Fig. 13 presents the cold side temperature response of the thermoelectric cooler for different pulse shapes. The minimum temperature is reached with the exponential rise pulse, because the input current is exponentially increasing and also the average pulse cooling power is lower (i.e. the area under the exponential curve (refer to Fig. 12) is lower than the ramp and square pulse). The ramp pulse can provide slightly higher temperature than the exponential
154
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Fig. 11. Variation of average cooling power, COP and average cold side temperature with different pulse width.
Fig. 12. Schematic of rectangular, ramp up and exponential rise pulse.
pulse and the square pulse have higher peak overshoot when compared with exponential and ramp pulse for the similar operating conditions because, the square pulse can cater more cooling load when compared with other pulse shapes. The COP and average cooling power of thermoelectric cooler for different pulse shape has been presented in Fig. 14. The average cooling power with the square pulse will be higher than the exponential and ramp pulse and therefore, its COP is also higher. This clearly shows the importance of pulse shaping in the modified pulse operation of the thermoelectric cooler. It can be concluded that the modified pulse operation of thermoelectric cooler with the square pulse can provide better COP and high average cooling
power when compared with other pulse inputs and normal operating conditions. The main advantage of the modified pulse operation of the thermoelectric cooler for building space cooling is the simultaneous improvement of the COP and cooling power by changing the operational scheme, which can be easily integrated with any available thermoelectric cooling systems. Moreover, the COP and cooling power can be further improved by integrating thermal energy storage and direct evaporative cooling techniques [11,12] with the modified pulse operation of thermoelectric cooler. This technique can be an added benefit to the applications which uses thermoelectric coolers.
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
155
Fig. 13. Cold side temperature of TEC for different pulse shapes.
Fig. 14. COP and average cooling power of TEC for different pulse shapes.
4. Conclusions The modified pulse operation of thermoelectric cooler for building space cooling has been studied with different pulse width, pulse current, different pulse cooling power conditions and with different pulse shapes. The following conclusions can be obtained from the study: For a typical operating condition, the modified pulse operation can improve the cooling power and COP of the thermoelectric cooler by 23.3% and 2.12% respectively when compared with the normal operation of the thermoelectric cooler.
The optimum pulse current ratio and the optimum pulse width for the modified pulse operation of the thermoelectric cooler has been found to be 2 and 5 s respectively. The square shaped pulse can provide higher cooling power and COP when compared to the ramp and exponential pulses. The modified pulse operation of the thermoelectric cooler for building cooling applications can provide better COP than the normal operating condition. This study can be helpful in the design of modified pulse operation of the thermoelectric cooler for building space cooling applications.
156
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Acknowledgements The author S.M. acknowledge the Indo US Science and Technology Forum (IUSSTF), Department of Science and Technology, Government of India for the Building Energy Efficiency Higher and Advanced Network (BHAVAN) fellowship, and the support from the Nano-enabled Energy Conversion, Storage and Thermal Management Systems Group (NEXT), Department of Mechanical Engineering, University of Colorado at Boulder, USA.
References [1] Kaushik SC. Solar refrigeration and space conditioning. Divyajyoti Prakashan; 1989. [2] Bansal PK, Martin A. Comparative study of vapour compression, thermoelectric and absorption refrigerators. Int J Energy Res 2000;24(2):93–107. [3] Riffat SB, Qiu G. Comparative investigation of thermoelectric air-conditioners versus vapour compression and absorption air-conditioners. Appl Therm Eng 2004;24(14):1979–93. [4] Hermes CJ, Barbosa JR. Thermodynamic comparison of Peltier, Stirling, and vapor compression portable coolers. Appl Energy 2012;91(1):51–8. [5] Cai Y, Liu D, Zhao FY, Tang JF. Performance analysis and assessment of thermoelectric micro cooler for electronic devices. Energy Convers Manage 2016;15(124):203–11. [6] Wang J, Zhao XJ, Cai YX, Zhang C, Bao WW. Experimental study on the thermal management of high-power LED headlight cooling device integrated with thermoelectric cooler package. Energy Convers Manage 2015;1(101):532–40. [7] Attar A, Lee H. Designing and testing the optimum design of automotive air-toair thermoelectric air conditioner (TEAC) system. Energy Convers Manage 2016;15(112):328–36. [8] Maneewan S, Tipsaenprom W, Lertsatitthanakorn C. Thermal comfort study of a compact thermoelectric air conditioner. J Electron Mater 2010;39 (9):1659–64. [9] Cosnier M, Fraisse G, Luo L. An experimental and numerical study of a thermoelectric air-cooling and air-heating system. Int J Refrig 2008;31 (6):1051–62. [10] Gillott M, Jiang L, Riffat S. An investigation of thermoelectric cooling devices for small-scale space conditioning applications in buildings. Int J Energy Res 2010;34(9):776–86. [11] Tipsaenporm W, Lertsatitthanakorn C, Bubphachot B, Rungsiyopas M, Soponronnarit S. Improvement of cooling performance of a compact thermoelectric air conditioner using a direct evaporative cooling system. J Electron Mater 2012;41(6):1186–92.
[12] Tan G, Zhao D. Study of a thermoelectric space cooling system integrated with phase change material. Appl Therm Eng 2015;5(86):187–98. [13] Irshad K, Habib K, Basrawi F, Thirumalaiswamy N, Saidur R, Saha BB. Thermal comfort study of a building equipped with thermoelectric air duct system for tropical climate. Appl Therm Eng 2015;5(91):1141–55. [14] Irshad K, Habib K, Thirumalaiswamy N, Saha BB. Performance analysis of a thermoelectric air duct system for energy-efficient buildings. Energy 2015;30 (91):1009–17. [15] Snyder GJ, Fleurial JP, Caillat T, Yang R, Chen G. Supercooling of Peltier cooler using a current pulse. J Appl Phys 2002;92(3):1564–9. [16] Yang R, Chen G, Kumar AR, Snyder GJ, Fleurial JP. Transient cooling of thermoelectric coolers and its applications for microdevices. Energy Convers Manage 2005;46(9):1407–21. [17] Ma M, Yu J, Chen J. An investigation on thermoelectric coolers operated with continuous current pulses. Energy Convers Manage 2015;1(98):275–81. [18] Shen LM, Xiao F, Chen HX, Wang SW. Numerical and experimental analysis of transient supercooling effect of voltage pulse on thermoelectric element. Int J Refrig 2012;35(4):1156–65. [19] Ma M, Yu J. Experimental study on transient cooling characteristics of a realistic thermoelectric module under a current pulse operation. Energy Convers Manage 2016;15(126):210–6. [20] Lv H, Wang XD, Meng JH, Wang TH, Yan WM. Enhancement of maximum temperature drop across thermoelectric cooler through two-stage design and transient supercooling effect. Appl Energy 2016;1(175):285–92. [21] Lv H, Wang XD, Wang TH, Meng JH. Optimal pulse current shape for transient supercooling of thermoelectric cooler. Energy 2015;1(83):788–96. [22] Shen L, Chen H, Xiao F, Yang Y, Wang S. The step-change cooling performance of miniature thermoelectric module for pulse laser. Energy Convers Manage 2014;30(80):39–45. [23] Manikandan S, Kaushik S. Transient thermal behavior of annular thermoelectric cooling system. J Electron Mater 2016:1–10. [24] Bywaters RP, Blum HA. The transient behavior of cascade thermoelectric heat pumps. Energy Convers 1970;10(4):193IN7195–4IN8200. [25] Cheng CH, Huang SY, Cheng TC. A three-dimensional theoretical model for predicting transient thermal behavior of thermoelectric coolers. Int J Heat Mass Transfer 2010;53(9):2001–11. [26] Ozisik N. Finite difference methods in heat transfer. CRC Press; 1994. [27] Lertsatitthanakorn C, Wiset L, Atthajariyakul S. Evaluation of the thermal comfort of a thermoelectric ceiling cooling panel (TE-CCP) system. J Electron Mater 2009;38(7):1472–7. [28] Wang CC, Hung CI, Chen WH. Design of heat sink for improving the performance of thermoelectric generator using two-stage optimization. Energy 2012;39(1):236–45. [29] Manikandan S, Kaushik SC, Anusuya K. Thermodynamic modelling and analysis of thermoelectric cooling system. In: International conference on energy efficient technologies for sustainability (ICEETS), 2016. IEEE; 2016. p. 685–93.
Contents lists available at ScienceDirect
Energy Conversion and Management journal homepage: www.elsevier.com/locate/enconman
Modified pulse operation of thermoelectric coolers for building cooling applications S. Manikandan a,⇑, S.C. Kaushik a, Ronggui Yang b a b
Centre for Energy Studies, Indian Institute of Technology Delhi, India Department of Mechanical Engineering, University of Colorado at Boulder, CO, USA
a r t i c l e
i n f o
Article history: Received 11 January 2017 Received in revised form 28 February 2017 Accepted 1 March 2017
Keywords: Thermoelectric cooler Pulse operation Building cooling
a b s t r a c t This paper presents a modified pulse operation of the thermoelectric cooler (TEC) for building space cooling application. In the normal pulse operation of a thermoelectric cooler, a current pulse is given as the input, but in the modified pulse operation, along with the pulse current, the hot side heat transfer coefficient is also pulsed. A numerical model for the thermoelectric cooling system is developed assuming one-dimensional unsteady state heat transfer having convective heat transfer boundary condition at its hot side and constant heat flux boundary condition at its cold side. Then the thermoelectric cooling system is studied with variable pulse current ratio, cooling load, variable pulse width and with dissimilar pulse shapes. The result shows that, for a typical operating condition with the modified pulse operation, the thermoelectric cooling system can provide an average cooling power of 600 W with the COP of 1.01, which are 23.3% and 2.12% higher than the normal mode of operation (i.e. without current pulse) respectively. This study also found that the rectangular-shaped pulse can provide higher average cooling power and COP when compared with the ramp and exponential pulse. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction The use of thermoelectric coolers for building space cooling can be an alternative option to the conventional vapour compression refrigeration (VCR) systems, since it does not use any refrigerants, it is compact and provides noiseless operation. However, its coefficient of performance (COP) is lower when compared with the conventional VCR systems due to the low figure of merit (Z) of thermoelectric materials [1]. The thermoelectric cooling systems have been compared with the conventional vapour compression cooling system, vapour absorption cooling system and Stirling cycle cooling systems and it was found that the COP of the thermoelectric cooler system is lower when compared with the other above mentioned cooling systems [2–4]. These studies also suggested that, if the figure of merit of the thermoelectric materials are improved from the present commercially-available bismuth telluride alloy based thermoelectric coolers, which has an average dimensionless figure of merit (ZT) of 0.7 to higher than 3. Then the COP and the cooling power of thermoelectric coolers can be improved significantly, with a great potential to replace the con-
⇑ Corresponding author. E-mail address: [email protected] (S. Manikandan). http://dx.doi.org/10.1016/j.enconman.2017.03.003 0196-8904/Ó 2017 Elsevier Ltd. All rights reserved.
ventional air-conditioning systems for building space cooling applications. The design of thermoelectric coolers using number of transfer units for electronic cooling has been proposed by Cai et al. [5]. Wang et al. [6] have studied the cooling of light emitting diode using a thermoelectric cooler package, and found that the thermoelectric device is more sensitive to the operating current. Attar and Lee [7] have theoretically studied and experimentally verified the thermoelectric cooler for automotive cooling application with the cooling power of 400 W and COP of 1.27. Maneewan et al. [8] have experimentally studied a compact thermoelectric air conditioner with the operating temperature of 301 K, cooling power of 29 W and COP of 0.34, and also estimated the payback time of this system to be 0.75 years. Cosnier et al. [9] have numerically and experimentally studied the thermoelectric cooling and heating systems and found that the COP of the thermoelectric cooler may vary from 1 to 0.3 when the temperature difference between its hot and cold side varies from 15 K to 30 K. Gillott et al. [10] have experimentally studied the thermoelectric air-conditioner with the average cold side temperature of 295 K, cooling load of 220 W and COP of 0.46. To improve the COP of thermoelectric cooling systems for building cooling applications, researchers have integrated various techniques such as thermal energy storage and evaporative cooling with the thermoelectric air-conditioning system. Tipsaenporm et al. [11] have used the direct evaporative cooler to supply cold
146
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Nomenclature h m A C D I K L M N P Q R T U
heat transfer coefficient (W m2 K1) mass (kg) area (m2) specific heat (J kg1 K1) density (kg m3) current (A) thermal conductance (W K1) length number of infinitesimal sections of the thermoelement number of thermoelectric cooler modules electrical power (W) heat (W) electrical resistance (X) temperature (K) overall heat transfer coefficient (W m2 K1)
Greek letters a seebeck coefficient (V K1) k thermal conductivity (W m1 K1) q electrical resistivity (X m) r electrical conductivity (S m1) D difference air to the hot side of the thermoelectric cooler to assist the heat removal. This technique has improved the cooling power of the thermoelectric cooler by 40.6% and the COP by 20.9%. However, the COP of this system was still low as 0.52 with the cooling load of 74.5 W. Tan and Zhao [12] have studied the integrated thermal energy storage-assisted thermoelectric cooling system where the phase change material is used to store the cold energy during the night and then that coldness is used in the day time to assist the heat rejection from the hot side of the thermoelectric cooler with average cold side temperature of 296 K. This technique have improved the COP of the thermoelectric cooler from 0.5 to 0.78 (56%). Even though, these techniques improved the COP of the thermoelectric cooler, the integration of evaporative cooler/ thermal energy storage makes the system bulky and adds to the operation complexity. The solar operation of the thermoelectric cooler for building cooling application with the cooling load of 589 W for a room sized 2.8 2.7 2.6 m3 has been studied by Irshad et al. [13,14]. This system has a COP of 0.68 and the cooling load of 498 W with the average cold space temperature of 298 K. The pulse operation of the thermoelectric cooler can provide cold side temperature lower than the cold side temperature at steady state conditions. When the pulse current is applied to the thermoelectric cooler, the cold side temperature drops immediately because of Peltier cooling and then it increases to a peak value and gets peak overshoot because of accumulation of Joule heat in the thermoelement and Fourier heat conduction through the thermoelements from hot side to the cold side. Then the peak overshoot in cold side temperature decays to its steady state value as shown in Fig. 2(b). The cold side temperature has such profile, because the reaction time of Peltier cooling in the junction is much faster than the Fourier heat conduction. Snyder et al. [15] have numerically and experimentally studied the pulse thermoelectric cooling and showed that this technique can provide low cold side temperature which is comparable to the cold side temperature of a two-stage thermoelectric cooler. Yang et al. [16] have derived the expressions for the time to reach the minimum temperature and the holding time with the pulse operation of thermoelectric cooler. This study also analysed the influence of pulse shape and the thermoelement with variable cross section area on the temperature profile of the thermoelectric cooler. Ma et al. [17] have studied
Subscripts a environment b base plate c cold side of TEC dt infinitesimal time dx infinitesimal length gen generation h hot side of TEC i node number in input m mean/average n n type material o reference out output p p type material t time C ceramic Superscripts n incremental time
the continuous pulse operation of the thermoelectric cooling system and found that the periodic super cooling can be obtained, if the TEC is operated in the continuous current pulse mode. Shen et al. [18] and several others [19–22] have theoretically studied and experimentally demonstrated the pulse operation of the thermoelectric coolers. All these studies on pulse operation of thermoelectric coolers are envisioned to make use of the lower cold side temperature for thermal management of electronics/lasers. 1.1. Proposed methodology The studies on pulse operation of thermoelectric coolers proved that it can provide excess cooling for the period of the pulse width. Therefore, if the excess cooling (which is at lower temperature) is utilized at relatively higher temperature (288 K) for the building cooling applications, the average cooling power can be improved which would result in the improved COP. It was found that the thermoelectric cooling systems for building cooling applications have relatively lower COP. It was also identified that, if the cooling power of the thermoelectric cooler increases, the COP will decrease, but, for building cooling applications, the COP and the cooling power are equally important. It can be argued that the cooling power of the thermoelectric cooling system can be increased by adding more number of thermoelectric modules. However, it should be noted that, the effect of adding more number of modules to improve the COP of the system is uncertain and it also increase the capital cost of the system. Therefore, in this paper a new operating technique of the thermoelectric cooler is proposed which can improve the COP and cooling power of the thermoelectric cooler simultaneously without adding more thermoelectric modules. This is the novelty of the present research. 2. Thermodynamic modelling for pulse operation of thermoelectric cooler Fig. 1 shows the schematic of the thermoelectric cooler system for the building space cooling application. The thermoelectric properties and the dimensions of the thermoelectric device are given in Table 1. Here thermoelectric properties of commercially available bismuth telluride (Bi2Te3) has been used in this study.
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
147
Fig. 1. Schematic representation of the thermoelectric cooler for building space cooling application.
Fig. 2. Input variables and cold side temperature profile of normal pulse operation (a and b) and modified pulse operation (c and d) of TEC.
The following assumptions have been made for the modelling and analysis in this work: (1) One dimensional unsteady state heat transfer along the length of the thermocouple has been followed.
(2) Averaged thermoelectric material properties have been used for this study. (3) No convective and radiative heat transfer from the sides of the thermocouples to the surroundings. (4) The thermal and electrical contact resistances are ignored.
148
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Table 1 Properties of the thermoelectric cooler [23]. Sl. no.
Material parameter 1
1 2 3 4 5 6 7 8 9 10 11 12 13 14
1
Cp, Cn (J kg K ) CC,c, CC,h (J kg1 K1) Cb,c, Cb,h (J kg1 K1) mC,c, mC,h (kg) mb,c, mb,h (kg) Dp, Dn (kg m3) Ta (K) kp + kn (W m1 K1) qp + qn (X m) ap an (V K1) Ap, An (m2) Lp, Ln (m) N Number of thermocouples per module
Value 200 419 400, 850 2.95 105 8.64 105 10922.08 300 3.363 2.10 105 420 106 12.25 106 4.25 103 24 127
temperature at an incremental time (T nþ1 ) is a function of its temo perature at its previous time step (T no ), temperature of its nearest
2.1. Finite difference modelling of the thermoelectric cooler The total number of thermoelectric cooler modules used in this study is 24 and each thermoelectric cooler has 127 thermocouples. To simplify the numerical modelling, equivalent stage method has been assumed [24]. The one-dimensional unsteady state heat transfer model of the thermoelectric cooler has been developed by applying the energy balance at its hot side, cold side and in a small cross section along the length of the thermoelement. Then the finite difference approximation is applied to the energy balance equations to obtain the temperature profile of the thermoelectric cooler. The energy balance at the cold side of the flat thermoelectric cooler can be given as by Cheng et al. [25]
ðmb;c C b;c þ mC;c C C;c Þ
dT c ¼ Q c þ Q k IaT c dt
ð1Þ
where mb,c and mC,c are the mass of base plate and ceramic layer at cold side respectively and Cb,c and CC,c are the specific heat of base plate and ceramic layer at the cold side respectively. The similar nomenclature have also been followed at the hot side of the thermoelectric cooler. The Fourier heat conduction term Qk in the above equation can be given by Eq. (2), Qc is the cooling load and a can be defined as the combined Seebeck coefficient (a = ap an).
Q k ¼ kA
dT dx
ð2Þ
The energy balance at the hot side of the thermoelectric cooler can be given as
ðmb;h C b;h þ mC;h C C;h Þ hAs ðT h T a Þ ¼ h
dT h ¼ IaT h Q k hAs ðT h T a Þ dt
As AðT h T a Þ ¼ UAðT h T a Þ A
ð3Þ ð4Þ
where As is the surface area of the hot side heat sink, A is the cross section area of thermoelectric couples (A = 24 thermoelectric cooler modules 127 thermocouples Area of single thermocouple), and U is the overall heat transfer coefficient at the hot side of the thermoelectric cooler. The energy balance in a small section of the thermoelement (n type or p type) using the 1-D unsteady state heat transfer modelling can be given as 2
d T 2
dx
þ
I2 q kA
2
¼
DC p dT dt k
and node numbers from i = 0 to i = M + 1, where, Δx can be defined as the ratio of length of the thermocouple (L) to the number of sections (M). Then, by applying the finite difference approximation in the Eqs. (1)–(4) by following simple implicit method, we obtain two equations for the cold and hot side boundary temperatures and M 1 number of equations for the nodal temperatures of the thermocouples, accounting to M + 1 number of finite difference equations. Simple implicit finite difference method has been used to develop the finite difference equations, since it is always stable and more accurate when compared with the simple explicit method. Further details about the finite difference technique, simple implicit method, and the techniques to solve the finite difference equations can be referred from Ozisik [26]. The cold side temperature at boundary node i = 0 can be calculated using Eq. (5). It can be seen from the Eq. (5) that the cold side
ð5Þ
The thermoelements in the thermoelectric cooler has been divided into M number of equal parts with equivalent length (Δx)
node at an incremental time (T nþ1 1 ) and also the boundary condition (Qc).The terms (mb;c C b;c þ mC;c C C;c ) in the Eq. (1) and (mb;h C b;h þ mC;h C C;h ) in Eq. (3) have been expressed as M c C c and Mh C h respectively in subsequent equations wherever applicable.
M c C c 2Dx 1 Ia2Dx 1 2 1 þ þ þ kADt kA Dx2 Dx2 Dx2 F Dt 2 M c C c 2Dx 1 1 þ T no þ þ T nþ1 1 kADt F Dt D x2 Dx2 Q c 2 Dx 1 þb ¼0 ð6Þ þ 2 kA Dx
T nþ1 o
It should be noted that the subscripts (i) of the temperature (T) in Eqs. (6)–(8) corresponds to the node number and superscripts (n, n + 1,. . .) corresponds to the time and incremental time respectively. The hot side temperature of the thermoelectric cooler at an incremental time (i.e. node number i = M) can be calculated using Eq. (6). It can be seen from the below equation that, the hot side temperature at an incremental time (T nþ1 M ) is a function of its temperature at previous time step (T nM ), temperature of its nearest node at an incremental time (T nþ1 M1 ), the hot side heat transfer coefficient and the environmental temperature. M h C h 2Dx 1 Ia2Dx 1 2 1 þ T nþ1 þ M kADt kA Dx2 Dx2 Dx2 F Dt hAs 2Dx 1 2 M h C h 2Dx 1 1 nþ1 n þ T þ T þ þ M1 M kA kADt F Dt Dx2 Dx2 Dx2 T a UA2Dx 1 þ þb ¼0 ð7Þ kA Dx2 For all the nodes other than the boundary nodes, the temperature distribution in the thermoelement (i.e., for i = 1 to M 1) can be given as
2 1 1 1 Tn nþ1 nþ1 þ T þ i þb¼0 þ þ T i1 iþ1 Dx2 F Dt Dx2 Dx2 F Dt
T nþ1 i
ð8Þ
The constants (b and F) mentioned in Eqs. (6)(8) can be given as follows:
b¼
I2 q kA
2
and F ¼
k DC p
The power input to the thermoelectric cooler can be calculated as
Pin ¼ aIðT h T c Þ þ I2
qL A
ð9Þ
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Then the COP of the thermoelectric cooling system considering the input power required by the fans at its hot and cold sides can be given as
COP ¼
Q c;av erage Pin þ P fan
ð10Þ
where Qc,average, Pin and Pfan can be defined as
Z tþDtpulse þtsteady 1 Q c ðtÞ dt; Dt pulse þ t steady t Z tþDtpulse þtsteady 1 Pin ðtÞ dt and Pin ¼ Dtpulse þ tsteady t Z tþDtpulse þtsteady 1 P fan ðtÞ dt Pfan ¼ Dt pulse þ t steady t
Q c;av erage ¼
where tsteady is the time taken to reach the steady state after the application of current pulse and Dtpulse is the duration of pulse width. The power input required by a fan has been assumed to be 14 W and a total 8 fans have been used in this study (4 each at the hot and cold side of the thermoelectric cooler). Here 4 fans are operated during the normal operation (without any pulse) to facilitate the heat rejection from the hot side of the thermoelectric cooler and to supply cold air to the cooling space respectively, and during the pulse operation (i.e. only for the period of Dtpulse), additional 4 fans (total of 8) have been operated to increase the cooling power and hot side heat transfer coefficient (refer Fig. 2(c)). 2.2. Modified pulse operation of thermoelectric cooler The difference between the normal pulse operation and the modified pulse operation of the thermoelectric cooler is shown in Fig. 2. In the normal pulse operation, a pulse current (Ipulse) large than the steady state current will be given for a period of time (Δtpulse) in addition to a steady state current (Isteady) after the thermoelectric cooler reached its steady state. When the current pulse is applied, the temperature of the cold side suddenly drops to a temperature lower than the temperature of the single stage thermoelectric cooler at steady state condition. It should be noted that, during the normal pulse operation, the hot side temperature of the thermoelectric cooler for the pulse period is higher than hot side temperature at steady state condition. The increase in hot side temperature may influence the cold side temperature of the thermoelectric cooler because of the excess Joule heat accumulated in the thermoelement for the pulse period and Fourier heat conduction through the thermoelement and this causes a higher peak overshoot in the cold side temperature, as shown in Fig. 2(b). Therefore, in the modified pulse operation, to reduce the peak overshoot in the cold side temperature, the hot side heat transfer coefficient should also be increased for a period of Δtpulse as shown in Fig. 2(c), which minimize the temperature rise in hot side of the thermoelectric cooler (during the pulse period) and therefore, reducing the peak overshoot in the cold side temperature. In building space cooling applications with thermoelectric coolers, the cold side temperature can be 288 K [11], so that the supply air to the cold space will be 296–298 K as per the thermal comfort level [27]. Therefore, in the modified pulse operation, the cooling power will be increased with the applied pulse current, so that the larger temperature drop in the pulse mode can be utilized to cater more cooling load, which increases the cold side temperature up to the required level for space cooling application. In practice, the pulse hot side heat transfer coefficient and the pulse cooling power can be achieved by increasing the mass flow rate of the fluid flowing through the hot and cold side heat exchangers of the thermoelectric cooler.
149
3. Results and discussion The cooling load requirement for a room sized 2.8 2.7 2.6 m3 with the thermoelectric cooling system has been evaluated to be 589 W [13]. To select the initial values for the operating current, the hot side heat transfer coefficient and the base cooling load, the thermoelectric cooling system has been analysed at steady state without any pulse current. The variation in steady state cold side temperature and COP of the thermoelectric cooler with the hot side heat transfer coefficient for a constant cooling load has been presented in Fig. 3. It shows that, with the increase in hot side heat transfer coefficient, the cold side temperature of the thermoelectric cooler decreases, this is because of the decrease in the hot side temperature of TEC. Thus, increasing the hot side heat transfer coefficient effectively reduces the temperature difference between the hot and cold side of the thermoelectric cooler (Th-Tc). After a certain value of hsteady, the rate of change of Th-Tc decreases, thus making the COP increase at smaller rate. Therefore, hsteady has been chosen as 200 W m2 K1, and this value agrees with Wang et al. [28]. The hot side heat transfer coefficient during the pulse mode (hpulse) is then assumed to be 2 times of hsteady. Fig. 4 shows the variation on cooling power and COP of a single thermoelectric cooler with different operating current (without any pulse) with hot side heat transfer coefficient fixed at 200 W m2 K1. With the increase in current, the COP of the thermoelectric cooler increases first to reach its maximum and then decreases as expected [29]. In the normal pulse operation, the thermoelectric cooler are operated at the current for maximum temperature difference between its hot and cold junction. In building cooling applications it is important to maintain the cold space temperature. Therefore, the operating current should be lower than the optimum current for maximum temperature difference. Hence, in this case, when the temperature difference between the hot and cold side of the thermoelectric cooler is 12 K, Isteady has been chosen as 4 A, with which the COP is maximum. With the operating current of 4 A, and the number of thermoelectric cooler modules of 24, the steady state cooling load can be 486 W.
3.1. Influence of pulse cooling load The variation in cold side temperature of thermoelectric cooler under different pulse cooling load with constant pulse width of 5 s is presented in Fig. 5. The pulse cooling load is the cooling load of the thermoelectric cooler during the current pulse period, which has been selected to meet the average cooling load requirement of 600 W. Thus, the pulse cooling load has been selected as 1850 W. Three points can be noted from Fig. 5: (i) as the Qc,pulse decreases, the cold side temperature during the period of current pulse reduces and vice versa if the pulse cooling load increases, (ii) as the pulse cooling load increases, the peak overshoot in the cold side temperature increases so that the average cold side temperature also increases, (iii) as the pulse cooling load increases, the time taken to reach the steady state increases. The cold side temperature of the thermoelectric cooler decreases with the decrease in pulse cooling load, because at higher currents thermoelectric cooler can cater more cooling load for a fixed cold side temperature. However, in this case, the operating current increases (pulse current) and the cooling load is low, which leads to the decrease in the cold side temperature. This statement can be analytically verified using Eq. (6). Moreover, the increase in the cooling load increases the temperature difference between the hot and cold side of the thermoelectric cooler, which results in increased Fourier heat conduction through the ther-
150
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Fig. 3. Variation of cold side temperature and COP of TEC with hot side heat transfer coefficient at steady state.
Fig. 4. Variation on cooling power and COP of TEC as a function of steady state current.
moelement and thus, increasing the peak overshoot in the cold side temperature at higher pulse cooling loads. If the peak overshoot is higher with constant heat transfer coefficient at the hot side, it takes longer time to reach back to the steady state. The average cooling power has been calculated by integrating the cooling power including the pulse cooling power for the period of 60 s (i.e. (t + Dtpulse + tsteady) (t) = 60 s). The time frame of 60 s have been used because, the cold side temperature reaches the steady state before 60 s. It can be argued that reducing the time frame may increase the average cooling load and COP, but by doing so, the average cold side temperature will increase than the required cold side temperature and also the cold side temperature may not reach the steady state temperature (after applying the current pulse). If the steady state temperature is not reached before applying second (or) consecutive current pulse, the average cold side temperature of the TEC will increase further and it is not
desirable. The average COP has been calculated as the ratio of cooling power (integrated over the period of 60 s) to the total input power integrated over the same period of time as mentioned in Eq. (10). Fig. 6 shows that the average cooling power and COP of the thermoelectric cooler increases with increase in the pulse cooling power. With the increase in pulse cooling power, the average cooling power and average cold side temperature increases, which results in the decrease in the temperature difference between the hot and cold sides and the increase in the COP. 3.2. Influence of pulse current ratio The influence of pulse current on the thermoelectric cooler has been studied with constant average cooling power and with the average cold side temperature, in the former the average cold side
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
151
Fig. 5. Cold side temperature of TEC with modified pulse operation (hsteady = 200 W m2 K1, hpulse = 400 W m2 K1, Isteady = 4 A, Ipulse = 8 A, Qc,steady = 486 W, and Δtpulse = 5 s).
temperature varies where in the later, the average cooling power varies. The pulse current ratio is defined as the ratio of the pulse current to the steady-state current. 3.2.1. Influence of pulse current ratio with constant average cooling power The cold side temperature of the thermoelectric cooler with different pulse current ratio for constant average cooling power of 600 W is presented in Fig. 7(a). It shows that the average cold side temperature decreases with the increase in pulse current ratio, which is because, at higher currents thermoelectric cooler can cater more cooling load. However, in this case, the operating current increases (pulse current) and the cooling load is constant, which leads to the decrease in the cold side temperature. The COP and average cold side temperature of the thermoelectric cooler for different pulse ratio with constant average cooling
load of 600 W are presented in Fig. 8. It can be seen that the COP decreases with the increase in pulse current. This can be explained with two reasons: (i) as the pulse current increases, the Joule heating in the thermoelements increases and thus, reduce the COP (ii) if the pulse current decreases, the average cold side temperature increases, which results in decreased temperature difference between the hot and cold side of the thermoelectric cooler and therefore, the COP of the thermoelectric cooler increases. Since the COP decreases linearly with the increase in pulse current ratio, the optimum pulse ratio can be selected based on the average cold side temperature to match the temperature requirement for building space cooling application. If the pulse current ratio is chosen below 2, the COP will increase, but the cold side temperature of TEC also increases and it in turn increase the cold space temperature. The optimum pulse ratio has been selected as 2. The COP of the thermoelectric cooler with the pulse current ratio of 2 is 1.01
Fig. 6. Average cooling power and COP as a function of pulse cooling load.
152
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Fig. 7. Cold side temperature vs. the pulse current ratio (a) with constant average cooling power (hsteady = 200 W m2 K1, hpulse = 400 W m2 K1, Isteady = 4 A, Qc,steady = 486 W, Qc,pulse = 1850 W, and Δtpulse = 5 s), (b) with constant average cold side temperature (hsteady = 200 W m2 K1, hpulse = 400 W m2 K1, Isteady = 4 A, Qc,steady = 486 W, and Δtpulse = 5 s).
Fig. 8. Variation in COP and the average cold side temperature of the thermoelectric cooler with pulse current ratio.
which is 2.12% higher than the COP of thermoelectric cooler under normal operating condition.
3.2.2. Influence of pulse current ratio with constant average cold side temperature The influence of pulse current ratio has also been studied with constant average cold side temperature of 288.6 K (i.e. the temperature difference between the average cold side temperature and the steady state cold side temperature is zero). It can be seen from Fig. 7(b) that with the increase in pulse current ratio the pulse cooling load should increase to keep the average cold side temperature constant. However, the cold side temperature oscillation increases which is not desirable for building space cooling application since it adversely affect the thermal comfort of the occupants. Therefore, it is not advised to operate the thermoelectric cooler with higher pulse current. Moreover, at the higher pulse current, the COP of the thermoelectric cooler decreases, even if the averaged cooling power increases as shown in Fig. 9.
3.3. Influence of pulse width The variation in cold side temperature of the thermoelectric cooler with different pulse width is presented in Fig.10. It can be seen that the peak overshoot in the cold side temperature of thermoelectric cooler increases if the pulse width increases. The hot side temperature of the thermoelectric cooler increases for the period of pulse width results in increased accumulation of Joule heat in the thermoelement and increases Fourier heat conduction through the thermoelement, thus increasing the peak overshoot. It should be noted that the minimum cold side temperature is constant, due to the constant current pulse ratio. It can be seen that the cold side temperature reaches its steady state before 60 s for all cases, but the increase in pulse width causes longer time to reach the steady state and the longer decay time is not desirable for building space cooling applications. The variation of the average cooling power, the COP and the average cold side temperature of the thermoelectric cooler with the pulse width is shown in Fig. 11. As the pulse width increases,
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
153
Fig. 9. Variation of average cooling power and COP of TEC with current pulse ratio and pulse cooling power.
Fig. 10. Cold side temperature of TEC for different pulse width (hsteady = 200 W m2 K1, hpulse = 400 W m2 K1, Isteady = 4 A, Ipulse = 8 A, Qc,steady = 486 W, and Qc,pulse = 1850 W).
the average cooling power and the average cold side temperature increase, which results in increased COP. Since the average cooling power, COP and average cold side temperature increase with the increase in pulse width, the optimum pulse width can be selected based on the acceptable average cold side temperature. The optimum pulse width has been found as 5 s even though, increasing the pulse width increases the COP and average cooling power. It should be noted that, with the increase in the pulse width, the average cold side temperature increases because of the increase in peak overshoot and it is not desirable. For a typical operating condition with pulse width of 5 s, the increase in average cooling power and COP are 23.3% and 2.12% respectively when compared with the normal mode of operation. 3.4. Influence of pulse shape The pulse shape effect on the modified pulse operation of thermoelectric cooler has also been studied. The influence of ramp up,
ramp down, exponential rise, exponential decay and square (rectangular) pulse on the cooling power and COP of the thermoelectric cooler have been studied. The schematic of ramp up, exponential rise and square pulse is presented in Fig. 12. The ramp down and exponential decay pulse can be the mirror image of the ramp up and exponential rise pulse with respect to the time axis. The pulse width for all the pulses has been maintained as 5 s and the square pulse has been used as the base case for the comparison. It should be noted that the cooling power and the heat transfer coefficient at the hot side follows the shape of the current pulse in the respective cases. Fig. 13 presents the cold side temperature response of the thermoelectric cooler for different pulse shapes. The minimum temperature is reached with the exponential rise pulse, because the input current is exponentially increasing and also the average pulse cooling power is lower (i.e. the area under the exponential curve (refer to Fig. 12) is lower than the ramp and square pulse). The ramp pulse can provide slightly higher temperature than the exponential
154
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Fig. 11. Variation of average cooling power, COP and average cold side temperature with different pulse width.
Fig. 12. Schematic of rectangular, ramp up and exponential rise pulse.
pulse and the square pulse have higher peak overshoot when compared with exponential and ramp pulse for the similar operating conditions because, the square pulse can cater more cooling load when compared with other pulse shapes. The COP and average cooling power of thermoelectric cooler for different pulse shape has been presented in Fig. 14. The average cooling power with the square pulse will be higher than the exponential and ramp pulse and therefore, its COP is also higher. This clearly shows the importance of pulse shaping in the modified pulse operation of the thermoelectric cooler. It can be concluded that the modified pulse operation of thermoelectric cooler with the square pulse can provide better COP and high average cooling
power when compared with other pulse inputs and normal operating conditions. The main advantage of the modified pulse operation of the thermoelectric cooler for building space cooling is the simultaneous improvement of the COP and cooling power by changing the operational scheme, which can be easily integrated with any available thermoelectric cooling systems. Moreover, the COP and cooling power can be further improved by integrating thermal energy storage and direct evaporative cooling techniques [11,12] with the modified pulse operation of thermoelectric cooler. This technique can be an added benefit to the applications which uses thermoelectric coolers.
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
155
Fig. 13. Cold side temperature of TEC for different pulse shapes.
Fig. 14. COP and average cooling power of TEC for different pulse shapes.
4. Conclusions The modified pulse operation of thermoelectric cooler for building space cooling has been studied with different pulse width, pulse current, different pulse cooling power conditions and with different pulse shapes. The following conclusions can be obtained from the study: For a typical operating condition, the modified pulse operation can improve the cooling power and COP of the thermoelectric cooler by 23.3% and 2.12% respectively when compared with the normal operation of the thermoelectric cooler.
The optimum pulse current ratio and the optimum pulse width for the modified pulse operation of the thermoelectric cooler has been found to be 2 and 5 s respectively. The square shaped pulse can provide higher cooling power and COP when compared to the ramp and exponential pulses. The modified pulse operation of the thermoelectric cooler for building cooling applications can provide better COP than the normal operating condition. This study can be helpful in the design of modified pulse operation of the thermoelectric cooler for building space cooling applications.
156
S. Manikandan et al. / Energy Conversion and Management 140 (2017) 145–156
Acknowledgements The author S.M. acknowledge the Indo US Science and Technology Forum (IUSSTF), Department of Science and Technology, Government of India for the Building Energy Efficiency Higher and Advanced Network (BHAVAN) fellowship, and the support from the Nano-enabled Energy Conversion, Storage and Thermal Management Systems Group (NEXT), Department of Mechanical Engineering, University of Colorado at Boulder, USA.
References [1] Kaushik SC. Solar refrigeration and space conditioning. Divyajyoti Prakashan; 1989. [2] Bansal PK, Martin A. Comparative study of vapour compression, thermoelectric and absorption refrigerators. Int J Energy Res 2000;24(2):93–107. [3] Riffat SB, Qiu G. Comparative investigation of thermoelectric air-conditioners versus vapour compression and absorption air-conditioners. Appl Therm Eng 2004;24(14):1979–93. [4] Hermes CJ, Barbosa JR. Thermodynamic comparison of Peltier, Stirling, and vapor compression portable coolers. Appl Energy 2012;91(1):51–8. [5] Cai Y, Liu D, Zhao FY, Tang JF. Performance analysis and assessment of thermoelectric micro cooler for electronic devices. Energy Convers Manage 2016;15(124):203–11. [6] Wang J, Zhao XJ, Cai YX, Zhang C, Bao WW. Experimental study on the thermal management of high-power LED headlight cooling device integrated with thermoelectric cooler package. Energy Convers Manage 2015;1(101):532–40. [7] Attar A, Lee H. Designing and testing the optimum design of automotive air-toair thermoelectric air conditioner (TEAC) system. Energy Convers Manage 2016;15(112):328–36. [8] Maneewan S, Tipsaenprom W, Lertsatitthanakorn C. Thermal comfort study of a compact thermoelectric air conditioner. J Electron Mater 2010;39 (9):1659–64. [9] Cosnier M, Fraisse G, Luo L. An experimental and numerical study of a thermoelectric air-cooling and air-heating system. Int J Refrig 2008;31 (6):1051–62. [10] Gillott M, Jiang L, Riffat S. An investigation of thermoelectric cooling devices for small-scale space conditioning applications in buildings. Int J Energy Res 2010;34(9):776–86. [11] Tipsaenporm W, Lertsatitthanakorn C, Bubphachot B, Rungsiyopas M, Soponronnarit S. Improvement of cooling performance of a compact thermoelectric air conditioner using a direct evaporative cooling system. J Electron Mater 2012;41(6):1186–92.
[12] Tan G, Zhao D. Study of a thermoelectric space cooling system integrated with phase change material. Appl Therm Eng 2015;5(86):187–98. [13] Irshad K, Habib K, Basrawi F, Thirumalaiswamy N, Saidur R, Saha BB. Thermal comfort study of a building equipped with thermoelectric air duct system for tropical climate. Appl Therm Eng 2015;5(91):1141–55. [14] Irshad K, Habib K, Thirumalaiswamy N, Saha BB. Performance analysis of a thermoelectric air duct system for energy-efficient buildings. Energy 2015;30 (91):1009–17. [15] Snyder GJ, Fleurial JP, Caillat T, Yang R, Chen G. Supercooling of Peltier cooler using a current pulse. J Appl Phys 2002;92(3):1564–9. [16] Yang R, Chen G, Kumar AR, Snyder GJ, Fleurial JP. Transient cooling of thermoelectric coolers and its applications for microdevices. Energy Convers Manage 2005;46(9):1407–21. [17] Ma M, Yu J, Chen J. An investigation on thermoelectric coolers operated with continuous current pulses. Energy Convers Manage 2015;1(98):275–81. [18] Shen LM, Xiao F, Chen HX, Wang SW. Numerical and experimental analysis of transient supercooling effect of voltage pulse on thermoelectric element. Int J Refrig 2012;35(4):1156–65. [19] Ma M, Yu J. Experimental study on transient cooling characteristics of a realistic thermoelectric module under a current pulse operation. Energy Convers Manage 2016;15(126):210–6. [20] Lv H, Wang XD, Meng JH, Wang TH, Yan WM. Enhancement of maximum temperature drop across thermoelectric cooler through two-stage design and transient supercooling effect. Appl Energy 2016;1(175):285–92. [21] Lv H, Wang XD, Wang TH, Meng JH. Optimal pulse current shape for transient supercooling of thermoelectric cooler. Energy 2015;1(83):788–96. [22] Shen L, Chen H, Xiao F, Yang Y, Wang S. The step-change cooling performance of miniature thermoelectric module for pulse laser. Energy Convers Manage 2014;30(80):39–45. [23] Manikandan S, Kaushik S. Transient thermal behavior of annular thermoelectric cooling system. J Electron Mater 2016:1–10. [24] Bywaters RP, Blum HA. The transient behavior of cascade thermoelectric heat pumps. Energy Convers 1970;10(4):193IN7195–4IN8200. [25] Cheng CH, Huang SY, Cheng TC. A three-dimensional theoretical model for predicting transient thermal behavior of thermoelectric coolers. Int J Heat Mass Transfer 2010;53(9):2001–11. [26] Ozisik N. Finite difference methods in heat transfer. CRC Press; 1994. [27] Lertsatitthanakorn C, Wiset L, Atthajariyakul S. Evaluation of the thermal comfort of a thermoelectric ceiling cooling panel (TE-CCP) system. J Electron Mater 2009;38(7):1472–7. [28] Wang CC, Hung CI, Chen WH. Design of heat sink for improving the performance of thermoelectric generator using two-stage optimization. Energy 2012;39(1):236–45. [29] Manikandan S, Kaushik SC, Anusuya K. Thermodynamic modelling and analysis of thermoelectric cooling system. In: International conference on energy efficient technologies for sustainability (ICEETS), 2016. IEEE; 2016. p. 685–93.