AbstractsEngineering

Energy harvesting becomes one of the main concerns of our modern society. A huge amount of energy is wasted in industrial plants as low grade heat. Thermoelectricity is a promising way to partly harvest this heat. However, three main challenges prevent its implementation in large-scale applications: (i) available thermoelectric (TE) materials are expensive and cannot be mass-produced; (ii) the classical thermoelectric generator (TEG) structure is not adapted to large-scale applications; (iii) design methods adapted to large-scale TE systems do not exist. The objectives of this thesis were to face these challenges. Firstly, several prototypes of TEG have been developed using a low-cost TE material processable at large-scale, the Fe2VAl compound. Their characterisation showed the importance of the joining technique on the contact resistances and its influence on the performances of a TEG. Secondly, a new planar TEG has been proposed. The experimental proof-of-concept of this TEG has been carried out and models have been developed for the understanding and the design of this TEG. Finally, a global design procedure for large-scale TE systems based on multi-objectives optimisation has been proposed and applied to design a TE heat harvesting system adapted to a galvanising steel line. The power density of Fe2VAl-based TEG has been improved by several orders of magnitude when using techniques involving Cu-Fe2VAl reaction. The planar TEG developed during this thesis shows performances similar to classical TEG but could be processed more easily and at a lower cost. The global design procedure overcomes existing procedures. Finally, the TE system designed with this method for a galvanising steel line could recover more than 150 kW of electrical power through the use of Fe2VAl. (FSA - Sciences de l)  – UCL, 2015