大象传媒

We use the state of the art machine learning methods to improve solar聽panel designs. Greve et al. have recently shown that adding聽nano-particles to solar panels can increase their efficiency. However,聽there are many possible configurations how the nano-particles can be聽added to the solar panels and time-consuming physics modeling is聽necessary to optimize the design. We will initially explore a聽2-dimensional parameter space of nano-particle configurations and design聽a machine learning model that predicts the light spectrum the solar聽panel can capture. Then we use this machine learning model to optimize聽the parameters. Extensions of the project will involve optimizing a more聽complex higher-dimensional parameter space.