실험계획과 머신러닝을 활용한 CNC 절삭공정 개선과 품질예측모델 개발 사례

Abstract

This paper presents a case study of systematically obtaining feature data and applying machine learning methods for a small CNC machining company that cannot obtain big data using sensors. In order to obtain the feature data, an experiment is planned and conducted using the 2-level 4-factor fractional factorial design with four machining process variables, and then 1) the outer diameter dimensional data is obtained using an automatic measurement tool and 2) appearance defects are visually inspected. An improved process conditions are determined to enhance productivity, to reduce tool wear, and to prevent defects. By analyzing the dimensional data and the number of non-defective/defective items obtained through observation, quality prediction models are also developed. This paper is expected to be used as a reference for small and medium-sized enterprises to improve the manufacturing processes in the future.