Método estruturado para integração do ChatGPT como consultor técnico em design for manufacturing aplicado à fundição de alumínio

Abstract

This dissertation aimed to demonstrate the feasibility and relevance of integrating Generative Artificial Intelligence (AI) tools into the Product Development Process (PDP) applied to the design of mechanical components, with an emphasis on Design for Manufacturing (DFM) applied on pressure die casting process. The literature review revealed significant gaps regarding the structured use of AI in industrial environments, despite its recognized advancements in automation and time optimization. In a context where repetitive tasks and frequent consultations with experts consume considerable resources, a method was proposed to position AI as a technical consultant assisting in the development of components manufactured by aluminum casting, providing fast, relevant, and customized recommendations during the design stages. The method was developed following the Design Science Research (DSR) approach and was applied in a case study involving a simplified model of a structural mounting bracket used to fix the transmission to the chassis, designed using the high-pressure die casting (HPDC) process. Five automotive industry engineers participated in a procedure that compared the component development process with and without the proposed method, assessing criteria such as accuracy, execution time, and user perception through questionnaires. The results indicated an average recommendation accuracy/utility rate of 88%, with an estimated 50% reduction in expert consultation time, in addition to the prevention of early-stage design errors and increased productivity. The main limitations identified were related to the reliability and usefulness of AI responses, the dependence on engineers’ experience for proper interpretation, the absence of testing in real industrial environments, and the limitations of AI communication interfaces. Overall, the research confirms that generative AI can serve as a strategic support tool in the PDP, offering significant gains in time and productivity. However, its effectiveness directly depends on critical analysis and the engineer’s expertise, reinforcing AI’s role as a complementary tool rather than a substitute.