REMAINING USEFUL LIFE PREDICTIONS FOR TURBOFAN ENGINE DEGRADATION SIMULATION USING DEEP ARCHITECTURE.

Posted on 22 Jun 2020 12:10 am

In many industries there is a growing awareness of reliability and durability of the manufacturing systems in the field of production of mechanical products. Health of the machineries and the time-series fatigue life of the products are monitored by the time-to-time service and maintenance of the products ensures there is less time-down of the turbofan engine breakdown and increasing the lifeline of the turbofan engine. With the technology of machine learning based traditional algorithms depicts the remaining useful life (RUL) of mechanical products which enables to check the life time of the product and the discussion of Deep Long-Short Term Memory (DLSTM), there is a major drop down of RUL prediction. Later, enhancing the DLSTM with both forward and backward direction will need to BiDLSTM (Bi-directional Deep Long-Short Term Memory). Lastly the study regarding attention based learn of Deep LSTM, with respect to time-series of prediction of RUL. All the experiments are carried out from an open source of NASA turbofans datasets.