Scientist Robot

This project will utilize data and expert knowledge-driven artificial intelligence algorithms, multi-scale physical simulation methods, and high-throughput automated experimental methods to conduct closed-loop material design, process optimization, mechanism elucidation, and application research for perovskite solar cells. The plan is to develop multi-channel, multi-variable automated intelligent experimental equipment for perovskites, enabling high-throughput perovskite preparation and high-dimensional characterization data collection. By establishing a large model intelligent agent and active learning optimization framework in the field of perovskites, the project aims to construct a comprehensive parameter association model and structure-function relationship spanning material formulation, preparation process, material property parameters, device structural property parameters, and device performance. Subsequently, the project will employ data-driven design to identify key materials, optimize formulation processes, and conduct application verification in single-junction, tandem, and large-area photovoltaic devices, ultimately achieving effective acceleration of performance enhancement and practical application of perovskite cells through artificial intelligence.

The project will address the following key scientific and technological issues: (1) Mechanisms for improving the inductive reasoning capabilities of AI agents in the field of material knowledge; (2) Guiding mechanisms for AI-accelerated design of key materials and process optimization based on multi-variable structure-function relationships; (3) Integration and analysis framework for multimodal, multi-source heterogeneous data streams; (4) Methods for enhancing automated throughput and data dimension collection in high-precision spatiotemporal control processes; (5) Design and fabrication of single-junction perovskite modules that achieve a synergy between high efficiency and high stability; (6) Development of key technologies and stability research for the preparation of large-area perovskite-silicon tandem devices.