Deep Learning Approach to Predict Remaining Useful Life of Axial Piston Pump

In the hydraulic machines like hydraulic pump, while in use, many faults start to appear in these machines resulting in the undesirable output. One of such parameters is the leakage fault, which is very common in pumps and motors. In pumps, it has been observed that over a period of time with increase in wear, leakage also increases. This relationship between the wear and leakage value is utilized here using the NARX NN to estimate the RUL of the pump. In this work, the different training algorithm has been used in order to train the NARX NN, and the models that has been trained are used to estimate the remaining useful life (RUL) of an axial pump that is being used to control the hydraulic system of sheet metal casting process. Results of the model are promising.