회전기기의 운전상태 분류를 위한 기계학습 알고리즘의 성능 비교

초록

PHM (Prognostics and Health Management) technology that uses sensor data to diagnose and predict machine fault is currently being applied in many fields. This study presents a case study of digital twin technology for fault diagnosis and fault detection of rotating machinery as a part of intelligent factory implementation aimed at proper operation rate. A study on the fault diagnosis for operating conditions of the rotating machinery was performed by applying the machine-learning algorithm to the bearing part of crane. Features were extracted using statistical parameters of the collected sensor data time domain and frequency domain, and the features were analyzed by correlation analysis. Also, based on the analyzed characteristics, the dimensional reduction technique was applied to visualize the operating state of the rotating machine, and the possibility of fault diagnosis for operating conditions was identified through the relevant examples.

키워드