APVC2021 Automated failure diagnostic system for a cylinder of the hydraulic pressing machine

Abstract

Due to the impact of the COVID-19 pandemic and declining working-age populations in certain countries, manufacturing and production sites have actively enhanced production efficiency and labor productivity by incorporating digital technologies. In light of these trends, we have developed a system aimed at swiftly assessing the operational status of machinery in order to optimize production efficiency within a manufacturing company, thereby reducing the time required for machine inspection and repair. To efficiently identify failures and anomalies in hydraulic presses, we have implemented a machine learning-based failure diagnostic system. Vibrational acceleration sensors were installed on the cylinder, the primary component of the press machine, and continuous signal data was collected. By establishing a model based on standard deviation, crest factor, and maximum signal values for normal operation, deviations and temporal changes in the data were utilized to detect failures and anomalies. Consequently, it is feasible to predict the occurrence of failures by monitoring these variations.