Table 1.

ML/DL methods used for PdM.

Goal	Learning Task	ML/DL Method	Data Source	Equipment/Process	Ref.
Failure Prediction (FP)	Anomaly Detection	Hierarchical Clustering	General faults	Time and Frequency	[81]
	Classification	RF, SVM and LR	Physical faults	Track geometry	[82]
		AE	General faults	Rolling bearing	[35,36]
				Spacecraft	[32]
				Transformers	[34]
				Rotor bearing systems	[37]
			Vibration	Tidal turbine	[33]
				Bearings	[29]
			Acoustic signals Sensor data	Motors	[83]
		DNN	Vibration	Bearings	[84]
				Gasoline engines	[54]
				Engines	[63]
			Vibration, pressure and speed	Diesel engines	[55]
			Optical and visual	Laser welding	[56]
		CNN	Vibration	Planetary gearbox	[65,68]
			Grinding faults	Abrasive belt wear	[66]
			General faults	Rotor bearing systems	[37]
			Vibration and images	Rotating machinery	[67]
		RNN	General faults	Rolling bearing	[70]
				Air compressor	[71]
				Air compressor in buses	[73]
			Sensor data	Turbofan engine degradation	[72]
			Time-frequencies	Cantilever beams	[75]
		GAN	Sensor data	Turbofan engine degradation	[79]
			Vibration	Induction motor	[78]
Remaining Useful Life (RUL)	Regression	Online-SVR	Vibration	Rolling Bearing	[85]
		PSO+SVM	Vibration	Rolling Bearing	[41]
		Bi-directional LSTM	Sensor data	Turbofan engine degradation	[42]
		AE	Acoustic signal Sensor data	Turbofan engine degradation	[38]
		CNN	Sensor data	Turbofan engine degradation	[10,69,86]
		RNN	Sensor data	Turbofan engine degradation	[74]