Determination of thermal emission spectra maximizing thermophotovoltaic performance using a genetic algorithm

TL;DR

This paper uses a genetic algorithm to optimize the thermal emission spectra of a radiator for thermophotovoltaic systems, considering thermal losses, resulting in significantly improved efficiency and power density.

Contribution

It introduces a coupled TPV model with a genetic algorithm to identify optimal emission spectra considering thermal effects, advancing radiator design for TPV systems.

Findings

Maximum efficiency of 38.8% achieved with a specific emission spectrum.

Maximum power density of 41708 W/m² achieved with a broader emission spectrum.

Optimized spectra outperform blackbody and tungsten radiators, feasible with metamaterials.

Abstract

Optimal radiator thermal emission spectra maximizing thermophotovoltaic (TPV) conversion efficiency and output power density are determined when temperature effects in the cell are considered. To do this, a framework is designed in which a TPV model that accounts for radiative, electrical and thermal losses is coupled with a genetic algorithm. The TPV device under study involves a spectrally selective radiator at a temperature of 2000 K, a gallium antimonide cell, and a cell thermal management system characterized by a fluid temperature and a heat transfer coefficient of 293 K and 600 Wm-2K-1. It is shown that a maximum conversion efficiency of 38.8% is achievable with an emission spectrum that has emissivity of unity between 0.719 eV and 0.763 eV and zero elsewhere. This optimal spectrum is less than half of the width of those when thermal losses are neglected. A maximum output power density of 41708 Wm-2 is achievable with a spectrum having emissivity values of unity between 0.684 eV and 1.082 eV and zero elsewhere when thermal losses are accounted for. These emission spectra are shown to greatly outperform blackbody and tungsten radiators, and could be obtained using artificial structures such as metamaterials or photonic crystals.