The Dynamic Performance And Thermo-electric Conversion Process On The Actual TEC System

For the continuous miniaturization and high packaging density of large-scale telecommunications and datacenter embedded processors, the interfacial thermal management is universal. Thus, there has been a considerable resurgence of interest for the all-solid-state thermoelectric cooling technology in cooling the active regions of high heat flux. However, there has no practical application with high thermoelectric figure of merit or new thermoelectric materials at present. So, how to optimize the cooling performance with the available material is still an important task with practical significance. More than that, the pulsed thermoelectric super-cooling behavior, namely a phenomenon of a large decrease in temperature instantaneously available for the interfacial heat dissipation, gradually reveals more advantages. But, most previous work on the transient super-cooling mainly focused on the minimum supercooling temperature achievable, and ignored to clarify quantitatively the extent of the interactional effects on the enhancement of the transient supercooling performance, as well as the dynamics analysis on the characteristic time related to the system thermostability.The research was supported by National Natural Science Fund, DDSF (Grant No.51246005), the Doctoral Scientific Fund Project of the Ministry of Education of China (Grant No.20120142110045), Chinese Academy of Sciences (CAS) through Open Project Fund on the Key Laboratory of Cryogenic Engineering (Grant No. CRYO201121), and Natural Science Foundation of Hubei Province (Grant No.2011CD13288). In this work, based on the existing thermoelectric conversion mechanism, we investigated a synthetic approach on analyzing the coupling effects under various boundary and initial conditions, as well as an exploratory work for the pulsed thermoelectric supercooling technology.Firstly, after systematically analyzed the comprehensive influence factors and the essential cooling characteristics during the steady-state thermo-electric conversion process, it can be found that TE module with a small cross section will not make the deterioration of cooling performance, but save money on materials. For a fixed working condition, the cooling capacity depends only on G factor, along with the best ratio interval of0.06cm-0.15cm. Besides, TE module of2-4stages can make a satisfactory cooling capacity.Secondly, the dynamic characteristics of an actual thermoelectric cooling system were investigated experimentally. When coupled with an additional thermal load attached to Peltier junctions, the Petier cooling effect can be enhanced remarkably in a short time scale, if raising the heat transfer rate of the cold-junction heat sink. Similarly, instead of fan and fin heat sink on the hot side of TEC, there is an increase of78%in COP (namely the value of0.48) for the TEC system, due to the development of a thermosyphon with two phases to dissipate heat from the hot side.Thirdly, an unsteady heat transfer model for the actual TEC system was established. Then, combined with eigenfunction method and finite difference method, a numerical solution expression was derived with time-term and space-term. Certainly, the experiments by the PWM based performance test also verified the feasibility of this numerical method.Lastly, with the input of user-defined pulse mode function, a revised unsteady model was established, involving time-dependent imposed voltage pulse and time-dependent thermal boundary conditions on the transient supercooling behavior as well as the response of characteristic time and the pulse operation parameters during the periods of pulse start-up, pulse-on time and pulse-off time. Then, the coupling interaction of the thermoelectric effects on the amount of the availably electrical conversion was described. With the monotonically increasing pulse shape, it is more appropriate to achieve the maximum supercooling capacity. Especially for the economic evaluation, sine voltage pulse shows a greater advantage over other pulse shapes. To make up the interfacial cooling losses when coupled with an additional thermal load, the appropriate increase of pulse time or pulse amplitude can contribute to the full use of the electrical conversion. From this work, it can be served as a theoretical basis to guide the process control and system optimization for a thermoelectric cooling system driven by pulse modes in future.