Longi Solar – Kechao Lu – 2025

Summary

Kechao Lu developed an AI-powered research platform to accelerate innovation in silicon solar cell technology at LONGi.

Goals

To accelerate innovation in the solar industry, LONGi enlisted EDF Climate Corps fellow Kechao Lu to tackle a critical R&D bottleneck. Researchers faced significant delays in finding specific, technical information within a vast body of scientific literature, slowing down the development of more efficient solar technology. Kechao’s mission was to design an intelligent platform that could understand and retrieve precise information, providing researchers with a tool to drastically shorten research and development cycles.

Solutions

Kechao executed a two-phase strategy to build a sophisticated question-answering system. First, he developed a foundational Retrieval-Augmented Generation (RAG) platform by processing over 60 core industry documents into a vectorized knowledge base for semantic search. To elevate the system’s accuracy and reasoning, he then constructed a domain-specific knowledge graph containing over 25,000 relationships between key technical concepts. This upgraded the platform to a more powerful GraphRAG model, for which he designed two novel hybrid retrieval methods to synthesize insights from both structured and unstructured data.

Potential Impact

The GraphRAG platform has the potential to significantly shorten R&D timelines for new solar technologies. By giving researchers a tool for rapid, accurate, and deep information retrieval, the platform can help accelerate the discovery of innovations that improve solar cell efficiency and reduce costs. This advancement in solar technology directly supports the global transition to cleaner energy, contributing to climate goals by making solar power a more accessible and powerful resource in the fight against climate change.