In continuous-process systems, failures of rolling-element bearings typically cause accidents, reduced productivity, and production-related financial losses. Therefore, predicting both the lifespan of rolling-element bearings and their replacement time is crucial for preventing machine system failures. Accordingly, numerous studies have reported various machine and deep learning classifiers for predicting the lifespan of bearings. However, these studies did not consider degradation trends of bearings. Thus, this study aimed to develop an algorithm to predict the lifespan of a bearing by considering its degradation trend. A vibration dataset of bearings was obtained at low and high speeds. Using a second-order curve-fitting model, various degradation patterns in the dataset were classified. Appropriate time-domain or frequency-domain feature variables applicable to the design of a classifier were determined according to classified patterns. In addition, the classifier was trained using multiple bidirectional long short-term memories. Finally, the performance of the developed classifier was verified experimentally.

Tool condition monitoring is one of the key issues in mechanical machining for efficient manufacturing of the parts in several industries. In this study, a tool condition monitoring system for milling was developed using a tri-axial accelerometer, a data acquisition, and signal processing module, and an alexnet as deep learning. Milling experiments were conducted on an aluminum 6061 workpiece. A three-axis accelerometer was installed on a spindle to collect vibration signals in three directions during milling. The image using time-domain, CWT, STFT represented the change in tool wear of X, Y axis directions. Alexnet was modified to learn images of the two directional vibration signals, to predict the tool condition. From an analysis of the results of learning based on the experimental data, the performance of the monitoring system could be significantly improved by the suitable selection of the data image method.