Deep Learning-based Data Augmentation for Hydraulic Condition Monitoring System
Kyutae Kim, Jongpil Jeong · Procedia Computer Science · 2020
It is very important to classify the anomaly data because most of data collected from the manufacturing plant are time-series data and can be analyzed for fault detection. Many researches have been conducted on deep learning over the last decades and it have shown good performance in solving many demanding classification problems. However most deep learning models require a lot of data for training because of the large number of parameters. When the amount of data is small, data augmentation can be a good solution. But data augmentation for hydraulic system has not been deeply studied yet. In this paper, a novel data augmentation is proposed to increase the amount of data for monitoring the condition of hydraulic system. Therefore, deep learning model based on our proposed method is applied to the classification of hydraulic system data and shows good performance in terms of accuracy and loss.