Periodic two-phase heat transfer coefficient in thermoelectric cooling mini evaporator

S. Filippeschi 0 E. Latrofa 0 G. Salvadori (corresponding author) 0 0 Department of Energetics Lorenzo Poggi, Pisa University , Via Diotisalvi 2, 56100, Pisa , Italy Highly compact Periodic Two-Phase Thermosyphon (PTPT) cooling devices joined with a thermoelectric cooler can allow a wide flexibility in the design of CFC-free refrigerators. In this paper a method has been presented to experimentally evaluate the PTPT evaporator heat transfer coefficient highly changing over time. In line with the development of CFC-free alternatives for refrigeration, thermoelectric cooling is widely regarded as the only pragmatic solution that could eventually replace vapour compression systems, realising sustainable environmental benefits. Thermoelectric refrigeration covers a wide range of applications [1]. Even though this technique has been mainly employed in small volume devices (as portable refrigerators or cold boxes [2]), in literature there are studies showing its capability to cool volumes of any size: from 40 litres, in the case of domestic refrigerators [3], to 3000 litres, pickup truck refrigerators [4], up to 220,000 litres, train carriage refrigerators [5]. The thermoelectric refrigeration is obtained when an electric current goes through one or more pairs of semiconductors (Peltier module). In a Peltier module the electrical energy is converted into a temperature gradient between the two junctions of semiconductors. The coefficient of performance (COP) of a thermoelectric refrigerator [6] is expressed by: - COP = v* where v* = conductors is: z = ( an + ap ) 2 ((g nln )1 2 + (g pl p )1 2 )2 Figure 2. Experimental facility. Experimental facility acquisition system (thermal resolution 0.1 K, accuracy 0.5 K, acquisition frequency 1/3 Hz) and stored in a personal computer. During the experimental activity, every test has been carried out with a constant mass of liquid inside the l.oop. The the.rmo-heater has been supplied by different values of electric powers Qe. For each Qe, different volumes VT (from 64 106 m3 down to 3 106 m3) have been tested. Each test starts as soon as the input electric power is supplied to the thermo-heater and it stops as a stable periodic regime is reached. At the test start, all the liquid volume VT is inside the evaporator and it is at environmental temperature 297 2 K. Periodic heat transfer coefficient observed are different in time: liquid convection, transition boiling, fully developed boiling condition and vapour convection. The periodic function, which represents the heat transfer coefficient in time, shows similar shapes for different VT, but it can change greatly in its maximum value. In a PTPT heat exchanger, applied to a thermoelectric cooler, the temperature oscillations at the junctions are deeply influenced by the periodic heat transfer coefficient function over time. A right design of the heat exchanger thermal capacity can damp the junction temperature oscillations and improve the COP of the cooler. A method to sharply measure the different heat transfer coefficient functions over time in a PTPT device, at different liquid volumes transferred every cycle, at different operative parameters and for different working fluids, must therefore be defined. Experimental results h(t) = Qd (t) Sd [TW (t) TS (t)] . where Qd(t) is the power dissipated from the surface Sd, and TW(t) and TS(t) are the temperatures of the copper dissipater wall and the saturated vapour, respectively. In these experiments, the te. mperatures TW and TS are measured over time (time increments 3 s). The power Qd(t) can be determined, at a generic time t, by the following thermal power balance: Figure 4. Tin1+1 Tin+1 Tin+1 Ti+n1+1 Ri1,i = Mi c (Tin+1 Tin ) t (5) = M1 c (T1n+1 T1n ) t where TWn+1 and Tn+1 are the temperatures experimentally measured at the time step S n + 1. All the energy balances, which are expressed by equations (5)(7), form a linear system of 10 equations with 9 unknown temperatures (T n+1 at the nodes 19) and 1 unknown heat transfer coefficient hn+1. The starting conditions of the system are Ti1 = TW1 1 i 9; h1 = Qe (Sd (TW1 TS1 )) Infrared thermographic detection of thermal power The bottom of the copper dissipater has been observed by an infrared thermocamera. The images acquisition frequency has been 30 Hz. This value allows 180 measurements of the temperature distribution during the time (6 s) as the evaporator is going to be empty, and the heat transfer coefficient abruptly decreases. On the other hand, in the same time only 34 measurements can be made by thermocouple. Unfortunately the heat transfer coefficient evaluation, from temperature measurements, results in an inverse heat conduction problem of very high complexity. The problem can be described by expression (9), by considering the axial symmetry: = Qe Se = Qd Sd for each infrared picture, 10 pixels on the right and 10 pixels on the left for every z level have been disregarded in the measurements. Moreover the 2 upper and the 2 lower pixels of the infrared picture are disregarded to eliminate the noisy effects which are introduced by the presence of the conical surface on the top and the thermo-heater on the bottom of the dissipater. The effective infrared picture area, which has been analysed during the test, is shown in Fig. 9. Within this picture, the temperature of each pixel at the same z level can be considered constant, with an error lower than 5 % with respect to the mean value. The thermal distribution, which is. measured with the infrared thermography, has been used to compute the heat flux Qm that is transferred over time from the bottom of the dissipater to the top, as shown in Fig. 9. The temperature distribution within the dissipater, in case of negligible he.at losses, can be considered one-dimensional time dependent and the heat flux Qm can be defined as: l z = Qm Se . Qm can be approximated by using the finite different equation (11) w 44 1 Tjn Tjn,y w 44 j=1 y =1 Figure 11. . Heat flux Q m: comparison of different evaluation methods. The heat transfer coefficients, calculated by the method described in previous section, are therefore confirmed by the thermographic analysis, even over those 24 seconds with the evaporator completely empty. The heat transfer coefficient functions can be used to design PTPT cooling devices joined with a thermoelectric cooler. Conclusions (...truncated)