138
Z. Wang et al.
AE Data
Fatigue Test
Start
End
Calculate the
Median
Standard
Deviation of
Noise
Floating
Threshold
Voltage
AE Event
AE Indexes
Compute the
Average
Fig. 13.2 Flow chart of improved acoustic emission index algorithm
In (13.1), x i is the No. i data value of analysed sample x. N is the length of the
analysed data. Symbol ‘median’ represents the median of a sample.
(2) Then, the noise threshold of each AE sample is estimated adaptively with
shrinkage method theory by L. Donohod [18], and use it as the floating threshold
voltage to replace the conventional fixed threshold voltage. Its mathematical
expression is:
T = γ × σ ×
2 lg(N )/N
(13.2)
In this Equation, γ is a threshold constant. It can be set according to the actual
noise level. σ is the estimation of noise standard deviation of data. N is the
length of data. T is the floating threshold voltage. It can adjust the voltage value
adaptively according to different sample conditions.
(3) In the end, each AE index is averaged. AE index obtained by floating threshold
voltage is averaged according to each sample with statistical indexes time
process accurately, and its data display can be reduced.
The flow chart of the improved AE index algorithm is shown as follows (Fig. 13.2):
Improved AE index can better describe comprehensive performance and characteristics of AE events with these steps above. It can provide a better fatigue damage
process information of ball bearing.
13.3 Test and Analysis
13.3.1 Ball Bearing Fatigue Test Rig
To verify effectiveness of the proposed algorithm, AE data is obtained from a selfdesigned accelerated fatigue test rig for ball bearings. The test bearing is a SKF-51115
thrust ball bearing. Hydraulic loading is used to stabilize the loading force. A GCr15
circular plate which has the same thickness as fixed bearing race with equivalent heat
treatment is used to accelerate the bearing fatigue process.
Apart from AE sensor (PAC), an acceleration sensor (PCB), an eddy current sensor
(XELI) and a K-type thermocouple (SWT) are used to make sure ball bearings state
can be detected better. The operating temperature, vibration and speed are measured
Z. Wang et al.
AE Data
Fatigue Test
Start
End
Calculate the
Median
Standard
Deviation of
Noise
Floating
Threshold
Voltage
AE Event
AE Indexes
Compute the
Average
Fig. 13.2 Flow chart of improved acoustic emission index algorithm
In (13.1), x i is the No. i data value of analysed sample x. N is the length of the
analysed data. Symbol ‘median’ represents the median of a sample.
(2) Then, the noise threshold of each AE sample is estimated adaptively with
shrinkage method theory by L. Donohod [18], and use it as the floating threshold
voltage to replace the conventional fixed threshold voltage. Its mathematical
expression is:
T = γ × σ ×
2 lg(N )/N
(13.2)
In this Equation, γ is a threshold constant. It can be set according to the actual
noise level. σ is the estimation of noise standard deviation of data. N is the
length of data. T is the floating threshold voltage. It can adjust the voltage value
adaptively according to different sample conditions.
(3) In the end, each AE index is averaged. AE index obtained by floating threshold
voltage is averaged according to each sample with statistical indexes time
process accurately, and its data display can be reduced.
The flow chart of the improved AE index algorithm is shown as follows (Fig. 13.2):
Improved AE index can better describe comprehensive performance and characteristics of AE events with these steps above. It can provide a better fatigue damage
process information of ball bearing.
13.3 Test and Analysis
13.3.1 Ball Bearing Fatigue Test Rig
To verify effectiveness of the proposed algorithm, AE data is obtained from a selfdesigned accelerated fatigue test rig for ball bearings. The test bearing is a SKF-51115
thrust ball bearing. Hydraulic loading is used to stabilize the loading force. A GCr15
circular plate which has the same thickness as fixed bearing race with equivalent heat
treatment is used to accelerate the bearing fatigue process.
Apart from AE sensor (PAC), an acceleration sensor (PCB), an eddy current sensor
(XELI) and a K-type thermocouple (SWT) are used to make sure ball bearings state
can be detected better. The operating temperature, vibration and speed are measured