Authors - Hector Rafael Morano Okuno Abstract - Nowadays, Large Language Models (LLMs) are utilized in various fields of knowledge to assist users in performing tasks after being prompted with a question. However, further exploration is needed to determine their scope, as they are constantly being trained to improve and enhance their capabilities. This article investigates an LLM to decide whether or not it can assist users in interpreting results from finite element analyses and whether it can suggest modifications to the analyzed parts to improve the results of said analyses. This work was conducted during the Cyber-Physical Systems course, held in the August-December 2024 semester, with students of mechatronics, robotics, and computational systems engineering. Among the results, it was found that LLMs are capable of recognizing images with beam diagrams featuring different types of support and loads, as well as correctly interpreting finite element analysis results generated by software tool applications, with the potential to provide recommendations for improving these results.
