13 Fatigue Evolution of Ball Bearing with Improved Acoustic …
141
0.3 mm. Hence, this stage can be defined as primary pitting stage. The cause of this
stage may be due to the damage accumulation leads to the gradual surface damage
expansion until the pitting is occurred. Trend amplitude increases gradually from
macro signal performance.
In D Area, amplitude in each index trend shows a sharp increase. And a high
level amplitude is maintained until the end. There are obvious and regular shock
components in AE signal at this stage. The impact is also significantly increased.
In this process, AE signal is released from material with high strain energy. At this
time, disassembly observation is carried out. The size of pitting has already reached
about 2.5 mm. The dramatic growth in AE signal is the rapid cracks growing on the
race surface. Therefore, this stage is defined as the spalling stage.
Among these improved AE indexesaveragege rise time and average duration have
a better consistency. And these two parameters can earlier excavation of the material
surface damage (dashed line) compare to other improved AE indexes. As early as
around 808.5 min, average duration and average rise time can detect damage on the
surface about 55 min earlier than the other indexes. AE event, average ring count
and total ring count have a good consistency. They have slight difference in trend in
B area. Compared with other parameters, it is more stable and less volatile.
13.3.3 Comparison and Verification
In order to observe the effect of the improved AE index. Conventional AE index
such as ring count, rise time, duration and amplitude parameter are brought in for
comparison. First, typical signals from each stage above are observed. The threshold
voltage is set manually to 0.0459 V (Consider the 20 dB preamplifier, the threshold
after conversion is 73.2363 dB), the analysis results of trend with fixed thresholds
are shown as follows:
From the figures below, conventional AE index with floating threshold is obviously
better than the trend of manual selected fixed threshold. However, the distinction
between each evolution stage is still not obvious. And its data is large and fluctuate.
It still not well matches with other statistical indexes in time. The fatigue process
accurate judgment of ball bearing is affected. By comparison, it can be known that
the data size with improved AE index is the same as the data size of the other statistic
index. It can be better matched with other statistical indexes in time and a better ball
bearing fatigue process information can be given. It is convenient for the subsequent
comprehensive judgment (Figs. 13.5, 13.6, 13.7 and 13.8).
In the end, trend analysis with 20 statistical indexes (Table 13.2) is also carried
out. Compared with the improved AE index, the standardized results can be divided
into three categories. See Figs. 13.9 and 13.10.
In Fig. 13.9, it can be seen that the trend of variance, RMS, Mean Square
Frequency, etc. These 11 statistic indexes are similar to each other. They are close
to the trend with the improved AE index. RMS Frequency can point out the slight
pitting moment better. However, it still has a certain lag with nearly 50 min compares
141
0.3 mm. Hence, this stage can be defined as primary pitting stage. The cause of this
stage may be due to the damage accumulation leads to the gradual surface damage
expansion until the pitting is occurred. Trend amplitude increases gradually from
macro signal performance.
In D Area, amplitude in each index trend shows a sharp increase. And a high
level amplitude is maintained until the end. There are obvious and regular shock
components in AE signal at this stage. The impact is also significantly increased.
In this process, AE signal is released from material with high strain energy. At this
time, disassembly observation is carried out. The size of pitting has already reached
about 2.5 mm. The dramatic growth in AE signal is the rapid cracks growing on the
race surface. Therefore, this stage is defined as the spalling stage.
Among these improved AE indexesaveragege rise time and average duration have
a better consistency. And these two parameters can earlier excavation of the material
surface damage (dashed line) compare to other improved AE indexes. As early as
around 808.5 min, average duration and average rise time can detect damage on the
surface about 55 min earlier than the other indexes. AE event, average ring count
and total ring count have a good consistency. They have slight difference in trend in
B area. Compared with other parameters, it is more stable and less volatile.
13.3.3 Comparison and Verification
In order to observe the effect of the improved AE index. Conventional AE index
such as ring count, rise time, duration and amplitude parameter are brought in for
comparison. First, typical signals from each stage above are observed. The threshold
voltage is set manually to 0.0459 V (Consider the 20 dB preamplifier, the threshold
after conversion is 73.2363 dB), the analysis results of trend with fixed thresholds
are shown as follows:
From the figures below, conventional AE index with floating threshold is obviously
better than the trend of manual selected fixed threshold. However, the distinction
between each evolution stage is still not obvious. And its data is large and fluctuate.
It still not well matches with other statistical indexes in time. The fatigue process
accurate judgment of ball bearing is affected. By comparison, it can be known that
the data size with improved AE index is the same as the data size of the other statistic
index. It can be better matched with other statistical indexes in time and a better ball
bearing fatigue process information can be given. It is convenient for the subsequent
comprehensive judgment (Figs. 13.5, 13.6, 13.7 and 13.8).
In the end, trend analysis with 20 statistical indexes (Table 13.2) is also carried
out. Compared with the improved AE index, the standardized results can be divided
into three categories. See Figs. 13.9 and 13.10.
In Fig. 13.9, it can be seen that the trend of variance, RMS, Mean Square
Frequency, etc. These 11 statistic indexes are similar to each other. They are close
to the trend with the improved AE index. RMS Frequency can point out the slight
pitting moment better. However, it still has a certain lag with nearly 50 min compares