디지털트윈 환경 기반 사출 코어금형 설계 자동화 기초연구

Abstract

Injection molds, composed of components such as upper and lower cores, mold bases, pins, and cooling channels, serve as the primary tooling for manufacturing plastic products. Despite the often simple geometry of molded products, the configuration and design of mold components remain highly complex, making the technical expertise and accumulated know-how of mold designers essential. However, the mold industry is facing increasing difficulties due to the discontinuation of academic programs dedicated to mold design, the aging of experienced designers, and the lack of incoming skilled personnel. To address these challenges, research on automating mold design has continued, and recent advancements in artificial intelligence (AI) have accelerated efforts to internalize expert knowledge through a variety of computational approaches. In this study, we conducted foundational research aimed at constructing a DT-AX platform capable of handling multiple domains by implementing and modularizing diverse processes within a digital-twin (DT) environment and integrating AI modules specialized for each process. Given the input dimensions of a bottle-cap model (diameter and height), the simplified outer dimensions of a core mold were predicted and subsequently used to generate a 3D model. The resulting STEP file was verified to be compatible with commercial CAD and simulation software. Overall, the results demonstrate the feasibility of implementing an automated mold-design module within a digital-twin environment. Future work will focus on diversifying design variables and increasing geometric complexity to develop modules that more closely approximate real-world mold design.