An intelligent Q&A system for vessel equipment faults based on large language models and retrieval-augmented generation technology

To address the low efficiency of vessel equipment fault diagnosis and the challenges in effective communication with commanders, this study proposes an intelligent solution. This solution employs natural language interaction to rapidly identify fault causes and recommend appropriate maintenance plans.

First, based on the domain-oriented design concept, an intelligent question answering system was developed by integrating large language models with retrieval-augmented generation technology. Then, a set of document preprocessing methods and comprehensive retrieval strategies were introduced to enhance system performance. Finally, a comprehensive evaluation scheme was devised to thoroughly assess the system.

Experimental results show that by merely using natural language to describe the observed fault symptoms, the system can accurately identify fault causes and provide corresponding maintenance solutions, significantly improving diagnostic efficiency. Compared to basic question answering systems, the optimized system achieved a twofold improvement in ROUGE score, a nearly 30% increase in BERTScore, and a 1.5-fold increase in expert ratings. Additionally, it reduced response time by 95% compared to traditional manual retrieval methods.

This offers robust technical support for the rapid restoration of equipment performance on coast guard vessels operating in complex mission environments, effectively enhancing their combat effectiveness and mission execution capabilities.