Abstract
Thermoelectric generators (TEGs) are compact and robust devices for converting heat into electrical power. In this work, we
investigate the response of a bismuth-telluride based TEG to the transient environment of a silicon production plant, where there
is a periodic change in the average temperature of the heat source. We establish a dynamic mathematical model that reproduces
results from industrial, on site experiments, both at steady-state and under transient conditions. By simultaneously changing the
design and location of the TEG, a peak power density of 1971 W=m2 can be obtained without exceeding material constraints of the
TEG, with an average power density of 146 W=m2. In the transient case, the average power density generated during one silicon
casting cycle is in all investigated cases found to be only 7 - 10% of the peak power density as the peak value of the power is only
maintained for a couple of minutes. The fractional area is defined as the ratio of the total area of thermoelectric modules to the
total system cross-sectional area of the TEG. We find that the power generated can be increased by reducing the fractional area,
provided that the TEG is at a fixed position. If the TEG can be placed as close as possible to the heat source without exceeding the
material constraints, the peak power density and the average power density reach maximum values as functions of the fractional
area, beyond which the power begins to decline. The optimal fractional area that gives maximum power depends strongly on the
cooling capacity. We find that with a higher cooling capacity, it is beneficial to design the TEG with a higher fractional area and
place it as close as possible to the silicon melt. Possible venues to improve the performance of TEGs that operate under transient
conditions are suggested
investigate the response of a bismuth-telluride based TEG to the transient environment of a silicon production plant, where there
is a periodic change in the average temperature of the heat source. We establish a dynamic mathematical model that reproduces
results from industrial, on site experiments, both at steady-state and under transient conditions. By simultaneously changing the
design and location of the TEG, a peak power density of 1971 W=m2 can be obtained without exceeding material constraints of the
TEG, with an average power density of 146 W=m2. In the transient case, the average power density generated during one silicon
casting cycle is in all investigated cases found to be only 7 - 10% of the peak power density as the peak value of the power is only
maintained for a couple of minutes. The fractional area is defined as the ratio of the total area of thermoelectric modules to the
total system cross-sectional area of the TEG. We find that the power generated can be increased by reducing the fractional area,
provided that the TEG is at a fixed position. If the TEG can be placed as close as possible to the heat source without exceeding the
material constraints, the peak power density and the average power density reach maximum values as functions of the fractional
area, beyond which the power begins to decline. The optimal fractional area that gives maximum power depends strongly on the
cooling capacity. We find that with a higher cooling capacity, it is beneficial to design the TEG with a higher fractional area and
place it as close as possible to the silicon melt. Possible venues to improve the performance of TEGs that operate under transient
conditions are suggested