Synthetic dataset to study the performance of perovskite solar cell simulations

<ns3:p>This paper presents a synthetic dataset to study the performance of perovskite solar cells (PSC) simulations using the simulation tool SCAPS-1D. The dataset consists of 18.570 simulated devices generated from four baseline device architectures and their respective photovoltaic performance values. The data set was generated through numerical simulations, and the evaluation of the electrical performance of the device was carried out by studying current density-voltage (J-V) curves under standard illumination conditions, temperature, and maximum applied voltage as working conditions, which were not modified. The dataset can be used to train different machine learning (ML) models using supervised methods or unsupervised techniques such as clustering or dimensionality reduction, which facilitate the identification of patterns or relationships between parameters. Thus, it can be useful in reverse design strategies to determine optimal configurations based on defined objectives. This work contributes to the development of PSC by providing a broad dataset for further analysis and optimization.</ns3:p>