Application of LLM Techniques for Data Insights in DHP
Jun Luo, Chao Ouyang, Yinggang Jing, Huajian Fang, Yunlong Xiao, Qing Zhang, Rongling Li · 2024
With the rapid development of China's nuclear power industry, data insights play an increasingly crucial role in enhancing the safety and operational efficiency of nuclear power plants. Traditional data analysis methods face challenges in handling massive, complex, and heterogeneous industrial data from multiple sources. The emergence of LLM (large language model) offers new solutions for industrial data insights. This article proposes a digital intelligent platform architecture for nuclear power LLM and thoroughly explores its applications in the nuclear power industry. The architecture includes data layer, model layer, analysis layer, and application layer, responsible respectively for data collection, storage, processing, analysis, and application. Finally, based on the architecture proposed in this article, prototypes for applications such as nuclear power industry statement generation and intelligent analysis assistant for patrol inspection data were developed, with a view towards future development directions.