디자인-생산공정 가공성 충돌 규칙 정립 및 AI 학습용 스키마 설계 연구

Manufacturability conflicts between industrial design intent and production-process constraints frequently lead to late design changes, tooling rework, and quality risks in distributed product development. This study defines a design–process manufacturability conflict as a discrepancy in which the intended aesthetic geometry cannot be realized under current manufacturing capabilities, or can be realized only at the expense of excessive cost, lead time, or quality risk. Focusing on four widely adopted processes in Korea, conflict cases and design guidelines were analyzed to derive representative rules. Each rule was organized by manufacturing process, conflict condition, and rationale, and the conditions were expressed as regular expressions using a limited set of patterns. Based on the regular expressions, an AI-ready schema was designed to store and manage manufacturability conflict rules in JSON format, enabling both rule-based checking and dataset construction for learning-based applications. The proposed framework supports early-stage verification of design data and provides a structured basis for accumulating and reusing manufacturability knowledge.