For better estimation of aero-engine RUL, we concatenate 1-D CNN and LSTM in a parallel structure. No matter which model you choose, you have to do data normalization first. Then convert data into the corresponding format, for example, [batch, time step, input dimension] is the input format of LSTM. These two pre-processing steps are very important.
migushu / rul-estimation-based-on-fcn-lstm Goto Github PK
View Code? Open in Web Editor NEWThis project forked from neovairis/rul-estimation-based-on-fcn-lstm
For better estimation of aero-engine RUL, we concatenate 1-D CNN and LSTM in a parallel structure.