Bearing failure diagnosis and prognostics modeling in plants for industrial purpose

Abstract When condition-based maintenance (CBM) is combined with proper decision support systems, it leads to enhanced utilization of resources and increased productivity which tends towards business efficiency. The forecasting the future condition, remaining operating life, or probability stable system behavior, based on data from acquired condition monitoring referred as prognosis an important part CBM process. Despite auto-regression integrated moving average (ARIMA) time series modeling, being long established dating back 1960s, has surged through new advances over years now recognized a major technique. Its application therefore investigated here in context FEMTO–ST Institute (Franche-Comté Électronique Mécanique Thermique et Optique-Sciences Technologies) bearing dataset. work discussed this article uses approach contributes modeling useful life (RUL) bearings plants, thereby helping prevent catastrophic failure before occurs. motivation for paper lies used structuring ARIMA models, adding value its by first ensuring stationarity signal using Dickey-Fuller Test, then makes easy accurate. result obtained compared results literature where neural network regression (NNR) was FEMTO competition. We checked contrasting our observations NNR well experimental National Aeronautics Space Administration (NASA)