At this point, the usage of Perovskite Solar Cells suffer from quite low efficiency scores as researchers search to improve the production process. However, traditional empirical trial and error methods for discovering and improving the quality of materials are unable to keep pace with the development of materials science in this fast-paced world, mostly due to their long development cycles and high costs. Thus, fueled by powerful data processing and high prediction performance, data mining and machine learning can leverage comprehensive expertise on materials science to reveal viable options for breaking the current ceiling on solar cells performance and increase their efficiency. Existing literature shows that efficiency is impacted by many environmental and functional factors. Thus, understanding the individual and collective impact of these factors requires complex analytics that are usually facilitated by computational models. We approach the problem of understanding and predicting the impact of various factors on the efficiency of solar cells as a two-step process:

(1) Regression models will be used to identify the hidden relationships and analyze the impact on the efficiency of the solar cells of functional parameters such as temperature, speed, distance of the nozzle, and pressure added after spraying.

(2) A convolutional neural network (CNN) model for computer vision will enable us to identify and classify deformations present within the cell structure in cross-sectional images taken with a microscope. Once these deformations are identified and quantified, they can be incorporated as an intermediary step into the regression model to establish the correlation between specific parameters and deformations. Finding these correlations is crucial to the success of predicting the optimal set of parameters that may lead to increased efficiency of the solar cells.

This project seeks to implement these machine learning techniques to the manufacturing process to reduce the time-consuming and costly nature while increasing the understanding of the interconnected nature of these environmental and and functional factors of the production process.