Optimization Study of Thermoelectric Energy System through Genetic Algorithms

• Main Authors: Yi-Hsiang Cheng, 鄭憶湘
• Other Authors: Chunkuan Shih
• Format: Others
• Language: en_US
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/53164723813308663757

博士 === 國立清華大學 === 工程與系統科學系 === 95 === This work presented a novel method based on genetic algorithms (GAs) to optimize thermoelectric energy systems. The objective of the optimization is on maximizing the cooling capacity or maximizing the coefficient of performance (COP) of thermoelectric cooling (TEC) systems. Two kinds of arrangements, including single- and two-stage TEC systems, have been studied. While optimizing a single-stage TEC system, structural parameters – including the thermocouple length, the thermocouple cross-section area and the number of thermocouple – were taken as the variables. While optimizing a two-stage TEC system, parameters – including the applied electrical current, the thermocouple length, the thermocouple cross-section area and the number of thermocouples – were considered. Two-stage TEC systems can be further categorized into three types, which are with two stages electrically connected in parallel, in series and in separate. A new mathematical modelling was also proposed to deal with the temperature-dependent material properties and to include the effects of contact and spreading thermal resistances between the two stages. For both single- and two-stage TEC systems, this study developed the design flowchart and programs that combine the mathematical modelling with GAs’ technique. All kinds of design constraints–space constraints and all others–can be considered and modeled during the optimization. The results indicate that the cooling capacity or COP can be increased by optimizing the parameters of TEC systems. This study also demonstrates that the new approach based on GAs can be used effectively to optimize the thermoelectric energy systems, and this method exhibits highly potential in handling complex design problems.