Data augmentation for remaining useful life estimation

In this paper, we introduce a new method with the goal of improving the precision of forecasts related to the remaining useful life (RUL) by utilizing data augmentation techniques. Estimating the RUL helps determine the number of cycles left before failure occurs, which enables proactive maintenance strategies. Deep neural models, such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and Transformers, have become widely popular in RUL prediction. However, the exploration of data augmentation techniques to enhance the accuracy of predictions in RUL research has been limited. Data augmentation is a commonly used technique in deep learning to increase the size and diversity of labeled training sets, which helps address overfitting and enhance performance. The main objective of this research is to bridge the gap in the literature by examining various time-series data augmentation techniques and analyzing their impact on the accuracy of RUL prediction models. The research primarily focuses on generating synthetic time-series data and evaluating the effects of different data augmentation techniques. To assess the accuracy of state-of-the-art RUL prediction models, experimental comparisons will be conducted with and without the inclusion of synthetic data.