Optimization of the thermophotovoltaic thermal emitter based on the machine learning

It is important to optimize the thermophotovoltaic (TPV) thermal emitter to enhance energy conversion. In this work, the optimization of the thermal emitter is achieved through changing the layer thicknesses and the material arrangement of the thermal emitter. The Transfer Matrix Method (TMM) and the Simulated Annealing Algorithm (SAA) are integrated to achieve majorization. By searching for the global optimal solution, the SAA can ultimately achieve an optimal multilayer thermal emitter structure with the value of figure of merit (FOM) = 85.34%, which is surpassing most previous works. The optimal thermal emitter is comprised of three materials: Si, SiO2, and W. The material W (105 nm) is as the base, and the optimal TPV thermal emitter structure is: Si (15 nm) SiO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> (35 nm) Si (35 nm) W (15 nm) Si (15 nm) W (24 nm) Si (35 nm) from top to bottom. After optimizing the structure, the work conducted a detailed analysis of the theoretical efficiency of TPV. Related properties reveal that the incident spectral energy of this thermal emitter is almost below the bandgap wavelength of photovoltaic cell. Upon evaluating the system efficiency of this TPV thermal emitter, it is found that the TPV thermal emitter exhibits better system performance.