Deep Learning-Based Predictive Maintenance Model for Air Cylinder in Manufacturing Systems

Abstract

With the rapid development of information technology and manufacturing automation, appropriate predictive maintenance for key devices and common components is important to ensure the safety, reliability, and productivity of manufacturing systems. Air cylinders are widely used in industrial robotic hands in automation systems, from assembly lines to mechanical manufacturing lines. However, the health status and life cycle of air cylinders can be affected by various factors during operations. This work proposes a fault classification model based on convolutional neural networks (CNNs) for various types of air cylinders. The feasibility of the proposed predictive maintenance strategy is validated on a battery cell pack test line.