We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
A comparative study on data pre-processing techniques for remaining useful life prediction of turbofan engines.
- Authors
Erdoğan, Meryem; Mercimek, Muharrem
- Abstract
This study delves into the application of Long Short-Term Memory (LSTM) for predicting Remaining Useful Life (RUL) in Turbofan Engines using the Jet Engine Simulated Dataset (C-MAPSS), systematically examining the combined impact of diverse data pre-processing techniques on RUL prediction, with a particular focus on the application of filtering and normalization. The initial filtering of the dataset employs Savitzky-Golay (SG), wavelet transform, and exponential moving average (EMA) techniques to effectively mitigate noise. Subsequently, minimum-maximum and zscore normalization techniques are implemented. Each filtering method, paired with distinct normalization approaches, is meticulously evaluated, and the performance of LSTM models in RUL prediction is assessed for each combination. The quantitative analysis of experimental outcomes indicates that normalization and filtering contribute to the improvement of the training phase in LSTM models, ultimately enhancing the accuracy of RUL prediction. The study emphasizes that the selection of an optimal data pre-processing structure plays a crucial role in influencing the efficiency of network training, underscoring the potential for optimizing RUL prediction through the application of the LSTM model.
- Subjects
TURBOFAN engines; REMAINING life assessment (Engineering); WAVELET transforms; PREDICTION models; MOVING average process
- Publication
International Journal of Materials & Engineering Technology (TIJMET), 2023, Vol 6, Issue 2, p50
- ISSN
2667-4033
- Publication type
Article