Banca de QUALIFICAÇÃO: FELIPE LIMA DE MEDEIROS

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : FELIPE LIMA DE MEDEIROS
DATE: 01/07/2025
TIME: 09:00
LOCAL: Meets: meet.google.com/ouz-xttd-zyw
TITLE:

Exploration of Structures in the Asset Administration Shell Pattern with a Conversational Agent Based on Large Language Models

KEY WORDS:

Asset Administration Shell, AAS, Industry 4.0, large language models

 

PAGES: 25
BIG AREA: Engenharias
AREA: Engenharia Elétrica
SUBÁREA: Eletrônica Industrial, Sistemas e Controles Eletrônicos
SPECIALTY: Automação Eletrônica de Processos Elétricos e Industriais
SUMMARY:

The Asset Administration Shell (AAS) is a standard for the digital representation of industrial assets, consolidating information and functionalities in a unified format to facilitate system integration. However, its technical complexity hinders access and exploration of these data by users without specialized
knowledge. This work proposes a conversational agent developed with the LangChain4j library, which connects the agent to large language models (LLMs) and can interpret natural language queries to retrieve information from AAS files. The solution aims to democratize information access and lower the entry barrier for new users. Experiments were conducted using three types of questions — direct, relational, and interpretative — applied to two groups of files with varying complexity levels. The models GPT-4o mini and GPT-4.1 mini were used to evaluate both response accuracy and execution time. Results demonstrate the effectiveness of the approach in information extraction, particularly in scenarios with fewer assets. Additionally, the proposed architecture is scalable and flexible, adapting to advances in language models and expanding application possibilities.

 

COMMITTEE MEMBERS:
Presidente - 1153006 - LUIZ AFFONSO HENDERSON GUEDES DE OLIVEIRA
Interno - 2885532 - IVANOVITCH MEDEIROS DANTAS DA SILVA
Externo ao Programa - 1775264 - GUSTAVO BEZERRA PAZ LEITAO - UFRN