Numerical Assessment of the Experimental Thermoelectric Cooling System Effectiveness

Solid-state thermoelectric elements, such as thermoelectric (TE) modules, can be used as cool ing devices. Small-sized TE modules, characterized by: lack of moving parts, no required refriger ants application as well as variable possible installation and operation positions, allow, in several speci fied cases, achieving the advantage in cooling process over the conventional refrigeration devices. This paper presents the results of the preliminary numerical determination of energetic efficiency of thermoelectric cooling system, applied for cooling a small-scale experimental room. The heat exchangers used in the cooling sys tem consisted of heat sinks and radiators installed on the both sides of the TE module. The numerical assessment included in this paper, based on a 3D model reflecting the experimental room and thermoelectric cooling system, allowed determining the relation between TE module power supply characteristics and cooling effects, as well as time-related temperature distribution inside the modeled experimental room. The commercial modeling software FLUENT, ANSYS 12.0 by ANSYS Inc. was applied in numerical calculations. The results of the performed labo ratory studies were used as a basis for model development, required input data, initial and boundary conditions. The results of laboratory tests showed the influence of amperage of power supply on the efficiency of cooling characteristics, as well as distribution of air temperature inside the experimental room. Calibration and validation of the developed model was also based on the results of laboratory experiment. The obtained results of numerical calculations showed the influence of amperage of power supply on efficiency of cooling characteristics as well as distribution of air temperature inside the experimental room.


INTRODUCTION
Nowadays, nearly 20% of the total electricity used in the buildings is related to indoor cooling. The International Energy Agency warns that without any action addressed toward increasing the energy efficiency of cooling process, energy demand will triple by the middle of the 21st century [IEA 2018].The use of thermoelectric technology in buildings has been investigated intensely for the last two decades. In many scientific publications, cited later in this work, TE modules are presented as promising devices that can be incorporated to many system configurations to achieve indoor thermal comfort and reduction of primary energy consumption, especially from the non-renewable resources. Thermoelectric cooling (TEC) is based on the Peltier effect, which causes absorption or rejection of the heat when an electric current flows through the interface of two semiconductor elements: n-type (N) or p-type (P), due to the difference in thermal energy transferred by the charge carriers [Enescu and Virjoghe 2014, Sarbu and Dorca 2018]. Modern and commercialized TE modules are built mainly from bismuth telluride (Bi 2 Te 3 ) (the information about different TE materials can be found in Selvan et al. [2019] and Soleimani et al. [2020]) and placed between two ceramic plates. The P-N or N-P junctions, forming a TE module, are connected electrically in series and thermally in parallel. The direction of the current flow determines which side is cold and which is hot [Twaha et al. 2016]. Schematic construction of the TE module is presented in Figure 1.   [Elarusi et al. 2017, Su et al. 2018 or even thermoregulatory garment [Zhao et al. 2018, Lou et al. 2020]. Owing to a compact size, fast reaction, elimination of moving parts and refrigerants application, opportunity to work also in heating mode and variable possible operation positions, TE modules can be successfully used as cooling devices and support the control of the personal thermal comfort parameters. In Fig. 2, four examples of TE implementation were presented. In the recent years, many scientific articles describing the use of TE modules to influence the indoor air temperature in buildings have been published in order to verify if they can be an alternative to more complex compressor chillers (despite the lower performance efficiency). For instance, two first cases refer to the TE modules used in active building envelopes. In Xu et al. [2007], an active building window (ABW) was presented in order to provide both, cooling and heating (Fig. 2c). In the long time tests, in the cooling mode, it was possible to reduce and maintain the temperature of the TE modules cold sides 5°C below the ambient temperature. In the heating mode, the temperature of the hot sides of TE modules was 40°C above the ambient temperature.
In Wang et al. [2021], a simple structure of the integral thermoelectric building wall (ITEW) was proposed and investigated. The authors prepared large-scale thermoelectric plates that can be used as building walls. Optimization showed that COP in the cooling mode ranged from 2 to 7 and the cooling capacity per unit area can range from 2 to 70 W/m 2 . Liu et al. [2015] presented the theoretical and experimental research on a thermoelectric solar air conditioner with the possibility of heat recovery for heating domestic water. The proposed system was investigated in three operating modes: space cooling, space cooling and heat recovery for water heating, space heating. The experiments showed that air cooling and water heating mode was characterized by relatively high COP coefficient, reaching slightly above 4.5. An experimental and simulation investigation of a thermoelectric air duct system (TE-AD) for space cooling under tropical weather conditions was described in Irshad et al. [2017]. In the presented case, TE-AD was supported by the photovoltaic (PV) system. The conducted experiments showed that, owing to this novel solution, 517 W of cooling capacity was provided to the test room and COP at level 1.15 was achieved. The authors estimated that in the particular case study, proposed system (combination of the TE-AD and PV system) could save 1806.75 kWh/year. Another example, presented in Arangure et al.
[2019], referred to a thermoelectric cooler-heat pump (TECHP) supported by heat pipes used as a heat exchangers (Fig. 2d). Together with waste heat recovery unit, TECHP was a part of a mechanical ventilation system applied into nearly zero energy building (nZEB). In the experimental results COP of TE cooler varied from 0.46 up to 5.25, depending on the cooling heat fluxes, mass flows and the position of the modules. As it may be observed based on the cited literature, TE modules can be applied in many different systems configurations supporting the cooling process. However, the application of the TE modules to a specific system should be additionally tested by fluid flow models to assess the effectiveness of the air and temperature distribution as the performance of the TE module is highly effected by the type of heat exchanger applied to support the heat transfer from the TE module ceramic surfaces [Liu et al. 2015, Afshari 2022 or even the geometry of the heat sink [Yilmazoglu 2016, Seo et al. 2018].
The purpose of the research presented in this paper was to numerically determine the efficiency of thermoelectric cooling system, consisting of heat sinks and radiators installed on the both sides of the TE module, installed in a small-scale experimental room. The performed numerical simulations, based on 3D model reflecting the experimental room and thermoelectric cooling system, allowed assessing the relation between TE module power supply characteristics and cooling effects as well as time-related temperature distribution inside the modeled experimental room.

MATERIALS AND METHODS
The 3D multivariate numerical modeling of thermoelectric cooling system efficiency presented in this paper was directly based on the laboratory measurements enabling to obtain the input data as well as the initial and boundary conations for numerical modeling and calibration of the developed three dimensional model. The laboratory tests were conducted using a laboratory set based on a small scale polystyrene chamber of internal cubature 0.125 m 3 , see Fig. 3. Thermal conductivity coefficient of the implemented Styrofoam material obtained from the available technical documentation, provided by the producer, was 0.034 W/(m·K) and the thickness of experimental chamber walls was 5 cm. The thermoelectric module QC-127-1.4-8.5MD by QUICK-COOL (Germany) was used in the experiment. The TE module was installed in movable cover of the chamber (the cover was also an insulation of the TE module). The tested module was applied to decrease the air temperature inside the experimental chamber. The heat generated on the hot side of the module was treated as a waste heat. On the both sides of the applied TE module, heat exchangers were installed: two sets of aluminum heat sinks with adjusted fans. Heat exchangers were crucial to improve the heat transfer between the two sides of the TE module, whereas fans improved the heat dissipation inside the chamber and waste heat dissipation to the surroundings.
The calibrated measurement and recording system was applied in order to measure the air temperature inside the chamber as well as temperature on the cold and hot side of the TE module. The measuring system consisted of six Pt 500 temperature sensors (measuring accuracy ± 0.1 K) connected to the data recorder. The thermoelectric module was powered by direct current (from laboratory power supply) of amperage equaled to 4 A, 5 A, 6 A and 7 A. One measurement series lasted 90 minutes. The temperature values in all points were measured every minute. Measuring series for one value of applied current were repeated three times.
Numerical calculations of heat transfer and temperature distribution inside the air experimental room, cooled by the tested TE module, were performed in the commercial finite elements modeling software FLUENT, ANSYS 12.0 by ANSYS Inc. The developed preliminary model, consisted of 70094 nodes and 345986 finite elements, reflected experimental room of volume 0.125 m 3 , radiator and fan. The developed geometry reflecting the real object of modeled domain is presented in Figure 4a while finite elements mesh, after discretization of the geometry, Numerical calculations were performed for four values of amperage supplying the TE module: 4 A, 5 A, 6 A and 7 A. The required input data covered air and aluminum characteristics, including density, specific heat, thermal conductivity and dynamic viscosity, assumed for preliminary model as based on three point linear characteristics. The assumed input data as well as initial conditions for air inside the experimental chamber are presented in Table 1.
The following characteristics of aluminum were assigned to the radiator zone: density 2719 kg/m 3 , specific heat 871 J/(kg·K) and thermal conductivity 202.4 W/(m·K). The assumed initial conditions covered the value of air temperature assigned to the air zone (see Table 1). There were two simplified types of boundary conditions assigned to the adopted model. The first one, the variable, time-related wall temperature in the radiator zone reflected outflow of heat from the modeled domain. The second, time-related wall heat transfer, calculated, using Eq. 2, for determined difference of temperature inside and  outside the experimental room, area of the wall and thickness (5 cm) as well as thermal conductivity of Styrofoam (0.034 W/(mK)). Both mentioned boundary conditions were determined for two periods, initial, with the duration of approx. 5 minutes, and remaining, the rest of experiment time duration. This partition was related to the temperature changes of radiator and air inside the chamber observed under laboratory conditions. The possible additional heat transfer through joints of Styrofoam plates was not included in the adopted model. The boundary conditions assigned to each variant of computation are presented in Table 2.

RESULTS
The time-related changes of temperature difference between cold and hot side of the TE module (ΔT) observed under laboratory conditions, directly affecting Qc, as well as the measured temperature of air inside the experimental room are presented in Figure 5. The results of temperature measurements presented in Figure 5 show the direct  Figure 5. Experimental results of measurements: air temperature inside the experimental room (T i ) and the temperature difference between cold and hot side of the TE module (ΔT) relation between the current amperage applied to power the tested TE module and obtained temperature values in monitored points. The higher the value of applied current, the greater the difference between temperatures on the cold and hot sides of TE module as well as the lower decrease in air temperature inside the experimental chamber were observed. Thus, the increase in current amperage triggers changes in cooling capacity or the total heat transfer rate of TE module, directly affected by the observed difference between measured temperatures of module cold and hot side. The increased temperature on the hot side, related to the increased amperage, in case of insufficient heat exchanger application, results in decreased cooling efficiency, reflected by the elevated air temperature inside the experimental room. The above-mentioned phenomenon is directly related to the changes in Peltier's and Joule's heat as well as heat transfer of TE module. Figure 6 shows the comparisons of two determined values: Q c -heat absorbed form the chamber during the experiments according to Eq. 3 as well the total energy consumed by the cooling system calculated as a sum of: E TE -energy consumed by the TE module and E fan -energy consumend by the fans in the heat exchangers. Again, according to the data presented in Fig. 6, the value of current amperage applied to power the studied TE module directly affects its cooling efficiency. The best cooling results, understood as Qc, were noted for the lowest applied current amperage i.e. 4 A and 5 A. For higher current values, a decrease in cooling system efficiency was observed. In contrast, the energy consumption by the system increases significantly with the increase in the amperage applied to the TE. It is connected with the energy delivered to the TE module as the energy consumed by the fans was the same for all tested series (approx. 8 Wh]. The achieved COP values (taking values from 0.01 to 0.14) are presented in Figure 7 and were calculated according to Eq. 4. The best results were achieved for the appied current value of 4 A and equaled 0.14. For the higher values of the cuurrent applied to the module, the decrease in the cooling performance can be observed, which is a result of the additional heat generated inside the module (Joule's heat) as well as the heat conduction of the heat form the hot side of TE module to the cold side of the module that was already mentioned in this work. The achieved values of COP are simmilar to ones presented in the literature for the simmilar systems [Afshari 2022]. Figures 8-11 show the distributions of modeled air temperature and velocity magnitude inside the experimental chamber for selected time steps of simulation duration achieved due to implementation of the Fluent software. The observed air velocity distributions are comparable in all studied cases. In contrast, the determined time related air temperature values and spatial distribution vary  depending on the applied current value. The highest ratio of modeled internal temperature values drops were noted for the lowest applied current value. In Table 3, the comparison of measured and modeled values of fi nal internal air temperature, obtained for diff erent amperage is presented. The diff erence between the obtained results for diff erent compared series varies from 0.37 to 0.76%, in

CONCLUSIONS
In this paper, the assessments of a thermoelectric cooling system operation as well as its effectiveness numerical determination were performed.  On the basis of the conducted experimental and computational analysis, the following conclusions can be drawn. The general relation between power characteristics and effective cooling energy rate, affected by the TE module total heat transfer rate and heat of thermal conduction from outside environment, was observed. Generally, increase in the value of current supplying the tested TE module equipped in the particular heat exchanger reduces its cooling efficiency reflected by relatively low decrease in air temperature inside the cooled room and significant increase in measured temperature on the hot side of the module.
The observed cooling effect depends directly on the current applied to the TE module. The best cooling results were achieved for the 4 A current, for which the temperature inside the experimental room decreased from 20.5 to 11°C. For this series, the efficiency of the system was also the highest -COP = 0.14. In general, application of TE module and the proposed heat exchanger allowed decreasing temperature inside the experimental room by 2.1-9.5°C. The heat obtained on the hot side (treated in this experiment as a waste heat), reflected by high values of the temperature determined on the hot side of TE module, can be reused for a certain purpose such as preliminary heating of domestic water. Thus, the overall system efficiency can be improved. The applied method of numerical modeling allowed successful assessment of time-related changes in velocity of flow and temperature distribution of air inside the experimental chamber.
Numerical calculations showed a relation between current amperage and local air temperature changes inside the experimental room. The highest decrease in temperature drop was related to decrease in the applied amperage value of TE module power supply. In all modeled variants, distribution of air velocity inside the modeled room was comparable, because it was closely related to fan efficiency, which was assumed the same for all studied cases.