Abstract
Fault diagnosis is essential for ensuring safety and efficiency in chemical process industries. Conventional diagnostic systems often generate raw numerical outputs that require extensive human interpretation, increasing the operator's workload and slowing decision-making during abnormal events. To overcome these limitations, this work introduces a model context protocol (MCP)-integrated fault diagnosis framework, where a Large Language Model (LLM) functions as the MCP client, coordinating multiple diagnostic tools through a unified protocol. Within the proposed framework, the LLM interacts with specialized diagnostic tools, including a convolutional neural network-based fault diagnosis model and an ensemble-based variant for uncertainty-aware analysis. The LLM synthesizes the outputs of these tools and generates operator-oriented natural-language reports that summarize diagnostic results and explicitly communicate uncertainty, thereby supporting more transparent and efficient decision-making.A benchmarking study based on the Tennessee Eastman (TE) process is conducted to evaluate the framework using multiple LLMs under identical settings. The evaluation assesses diagnostic accuracy, tool adherence, and operator report quality using an LLM-as-a-judge methodology. Experimental results show that LLM performance varies significantly across models, and that reliable tool calling and uncertainty-aware reasoning are more critical than model size for effective fault diagnosis assistance.Overall, the proposed framework demonstrates strong potential for enhancing fault diagnosis workflows in chemical processes by improving usability, interpretability, and decision-making efficiency without modifying existing diagnostic algorithms.