Thermoelectric generators (TEGs) are solid state devices that are capable of converting heat into electricity. These devices take advantage of the Seebeck and Peltier effects that describe how a voltage and current are created between two dissimilar metals, semiconductors or other materials. In recent years, considerable attention has been made to developing new materials for the use in thermoelectric generators because they have the potential to increase energy efficiency through the recovery of waste heat and for use in remote areas. The performance of thermoelectric materials is often measured by the dimensionless figure of merit. The figure of merit is essentially the ratio of electrical power output to the heat input. Currently, most TEGs that are commercially available have a figure of merit of one but emerging materials promise figure of merits closer to three which is believed to be the number required to make these devices commercially viable. Further, these devices contain no moving parts which make them very reliable since they do not fatigue. With the introduction of more advanced and efficient material coupled with the reliability they bring, TEGs may offer a serious alternative to existing power generation technologies in the future.
However, to effectively design applications that use thermoelectric devices it is necessary to accurately model the performance of these devices over a range of conditions. This thesis presents a thermoelectric generator model that has been validated experimentally. The model developed in this thesis is able to provide the voltage, electrical current, electrical power outputs and thermal efficiency of a device. For its inputs, the surface temperatures of a heat source and cold heat sink that provide heat flow through the TEG are required along with the electrical load that is applied to the device. From these inputs the model calculates the heat transfer coefficient of the thermal circuit between the heat source and sink to the surface of the TEGs to estimate the temperature gradient that is formed across the device in order to calculate the outputs of the model.
To experimentally validate the model, an experimental rig with a HZ-9 thermoelectric generator from Hi-Z was sandwiched between ceramic wafers and two heat exchangers with thermal grease applied between all of the surfaces to improve thermal conductivity. Attached to each heat exchanger were two type-K thermocouples that measured the temperature change of the fluid across the heat exchanger, a volumetric flow meter, and a Julabo recirculation heater/refrigeration unit that could be used to heat or cool the fluid. The hot side of the
thermal-fluid circuit contained ethylene glycol that was heated up to 165 oC and the cold side circuit used water at temperatures between 5 oC and 70 oC. The purpose of the heat exchangers was to provide heat to the system at different temperature gradients and to measure the thermal power consumed by the TEG to generate its electrical power. To determine the electrical outputs of the TEG, a multimeter was used to measure the open circuit voltage of the device with no load attached and the voltage drop across the external load when a load was attached to imply current using Ohms law. This current was then used to calculate the power dissipated by the load. For each temperature gradient a range of electrical loads were applied to verify the peak power as predicted by the maximum power transfer theorem.
Using this experimental apparatus, the electrical outputs of the model were validated but the thermal efficiency was not due to issues with measuring the volumetric flow rate of the fluid circuits. However, based upon the voltage, electrical current and electrical power output of the model over different temperature gradients and average module temperature, several observations can be made that can assist a designer in developing applications using thermoelectric modules
Firstly, the open circuit voltage, electrical current and peak power outputs of a device all increase when the temperature gradient across the device becomes larger. Further, the peak
power of the device occurs when the applied electrical load is matched to the internal resistance of the module. Secondly, for a fixed temperature gradient, an increase in the average temperature across the TEG will cause an increase in the open circuit voltage of the device and a reduction in the peak power output. Additionally, the electrical load that gives peak power also increases with the average temperature. Based upon the finding of this thesis, the most optimum scenario in employing these devices would be to create the largest temperature difference possible across the TEG but lower the average temperature of this difference.
To conclude, this thesis has provided a model that has had its electrical outputs validated experimentally. This model can be used to assist a designer in developing applications that use thermoelectric technology. Further, the model provides an insight into the optimum way to employ these devices.