A Novel Clustering Method Combining ART with Yu’s Norm for Fault Diagnosis of Bearings

Table 1

Time-domain and frequency-domain feature parameters.

⁢Time-domain feature parameters		⁢Frequency-domain feature parameters
Feature	Equation	Feature	Equation
Mean ()		Mean frequency
Root mean square ()		Average frequency that is wave shape of signal crosses the mean of time-domain signal
Standard deviation ()		Stabilization factor of wave shape
Skewness ()		Coefficient of variability
Kurtosis ()	−	Frequency-domain skewness
Crest factor (CF)		Frequency-domain kurtosis
Latitude factor (LF)		Root-mean-square ratio
Shape factor (SF)		⁢where is power spectrum density which is obtained by the Welch method. is the number of spectrum lines used to calculate the parameters, and⁢,
Impulse factor (IF)
where is a signal series for , and is the number of data points.