Simulation Of Multistage Cascaded Thermoelectric Generators With Adaptive Finite Element Method

A thermoelectric generator (TEG) is a device to convert the thermal energy to electrical energy using the Seebeck effect. The more cascaded or more thermocouples in TEG, the more efficient the device is in terms of producing voltage and power. Modelling efforts have been done by many researchers using different approaches including a full-field analysis with the finite element method, but full analysis of a cascaded model is not present in the literature. One of the main reasons is that accurate simulation of cascaded TEG is hard to achieve due to the nonlinearity of thermoelectric effect. For this purpose, this project aims to construct a full analysis of TEG using the adaptive finite element method (FEM) as it is generally accepted to attain highly accurate results with nonlinear effects. Various geometrical models with different level of complexity are considered in the study, where the finite element models are solved using linear and non-linear solvers combined with an hp-adaptivity. The results are visualised and post-processed to assess the accuracy. The major outcome for the project is the assessment of the accuracy of in-house adaptive FEM code used by comparing the power output from the simulation with the theoretical prediction. Although the results obtained for temperature are mostly accurate, the accuracy of voltage potential and power output varies according to geometrical complexities. For example, the voltage for single thermocouple is 0.033V while the voltage for 8 sets of cascaded TEG is 0.44V. In general, a more complex design with more thermocouples produces a higher value of maximum voltage and power output, which is consistent with the theory for cascaded TEG. In this respect, the main objective of the project has been achieved