디버링 공정의 효율성 향상을 위한 로봇 기반 디버링 자동화 시스템 개발
Development of a Robot‑Based Automated Deburring System to Improve Deburring Process Efficiency
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
This study converts the manual deburring and inspection workflow for die‑cast components into a robot‑based automated cell and develops a vision‑driven burr detection and path‑generation algorithm to verify field applicability. The proposed system comprises one transfer robot, two deburring robots, one vision robot, rotary jigs, conveyors, and PLC‑integrated control. A point‑cloud processing pipeline using a laser scanner was designed, including preprocessing, noise removal, contour extraction, and shape‑based burr classification. By comparing CAD‑derived paths with measured data, the system automatically identifies burr locations and sizes and adjusts tool RPM and deburring trajectories in staged levels according to burr magnitude. The deburring system supports both standalone teach‑in mode and real‑time vision mode, enabling handling of small/low‑frequency burrs as well as large, randomly distributed defects. Field implementation showed that a single deburring robot achieved a relative performance index of approximately 1.3 compared with manual operators (0.8), confirming efficiency gains through cycle‑time reduction. Further performance improvements are anticipated through high‑speed image‑processing optimization and coordinated multi-robot operation.
Abstract
This study converts the manual deburring and inspection workflow for die‑cast components into a robot‑based automated cell and develops a vision‑driven burr detection and path‑generation algorithm to verify field applicability. The proposed system comprises one transfer robot, two deburring robots, one vision robot, rotary jigs, conveyors, and PLC‑integrated control. A point‑cloud processing pipeline using a laser scanner was designed, including preprocessing, noise removal, contour extraction, and shape‑based burr classification. By comparing CAD‑derived paths with measured data, the system automatically identifies burr locations and sizes and adjusts tool RPM and deburring trajectories in staged levels according to burr magnitude. The deburring system supports both standalone teach‑in mode and real‑time vision mode, enabling handling of small/low‑frequency burrs as well as large, randomly distributed defects. Field implementation showed that a single deburring robot achieved a relative performance index of approximately 1.3 compared with manual operators (0.8), confirming efficiency gains through cycle‑time reduction. Further performance improvements are anticipated through high‑speed image‑processing optimization and coordinated multi-robot operation.