Big data-driven optimization framework for solar cell design

This work presents a simulation-driven, constraint-aware optimization framework for the systematic design of crystalline silicon solar cells. The proposed framework integrates automated large-scale device simulation with explicit feasibility filtering and objective-function evaluation to identify optimal design configurations within a predefined parameter space. A high-resolution simulation dataset comprising 14,641 design cases was generated using PC1D to capture performance trends with respect to key structural and electrical parameters. The optimal configuration identified through the proposed workflow achieved a conversion efficiency of 29.39% under the specified simulation conditions. To assess robustness, a subset of corresponding cases was independently evaluated using SCAPS, demonstrating consistent convergence to the same optimal design and confirming trend-level agreement across different simulation environments. It is emphasized that the proposed framework is demonstrated and validated exclusively for crystalline silicon solar cells in this study. The reported performance values represent deterministic simulation outcomes dependent on simulator assumptions, and experimental fabrication-level validation is required for practical deployment. The term “large-scale dataset” refers to a high-resolution simulation-driven design-space exploration rather than a machine-learning-scale dataset. Accordingly, the framework should be interpreted as a decision-support and trend-based optimization tool that can guide device-level design prior to fabrication, rather than as an absolute predictor of real-world performance or a turnkey solution for immediate deployment.