A Time-Series Data Generation Method to Predict Remaining Useful Life

Accurate predictions of remaining useful life (RUL) equipment using machine learning (ML) or deep (DL) models that collect data until the fails are crucial for maintenance scheduling. Because unavailable fails, collecting sufficient to train a model without overfitting can be challenging. Here, we propose method generating time-series RUL resolve problems posed by insufficient data. The proposed converts every training time series into sequence alphabetical strings symbolic aggregate approximation and identifies occurrence patterns in converted sequences. then generates new inversely transforms it series. Experiments with various prediction datasets ML/DL verified data-generation help avoid model.