PBSE Data Initialization Framework and Practice by Using LLM

With the increasing complexity of commercial aircraft and the rapidly changing market demands, the system engineering development pattern extensively adopted by aircraft OEM has evolved from the traditional document-based systems engineering (DBSE) to model-based systems engineering (MBSE) and pattern-based systems engineering (PBSE). MBSE employs models to describe products, while PBSE builds upon MBSE by utilizing engineering patterns, which are validated in advance, to enhance the efficiency and quality of data production in both MBSE and DBSE. However, during PBSE engineering practices, we have observed certain challenges, such as the barriers to initializing product S* models and the low efficiency in generating instances. Artificial intelligence for systems engineering (AI4SE) is an emerging concept aimed at creating a more efficient and user-friendly systems engineering implementation environment through the integration of artificial intelligence (AI), machine learning (ML), and related technologies. This paper explores the application of AI4SE in real-world engineering projects by leveraging large language models (LLMs) to develop a methodology that reduces the deployment threshold of PBSE for enterprises and enhances the efficiency of instance generation.