activity of Fig. 3, it is clear that the steady emissions could
be distinguished from the inordinate emission bursts.
Fig. 4(a) and (b) represents the correlation plots of event
counts versus peak amplitude for both genuine emissions
and inordinate emission bursts collected during the 90±
100% load sequence. In Fig. 4(a), almost all genuine emis-
sions fell into the 50±60 dB peak amplitude range with
below 20 counts value. The inordinate high density emis-
sion bursts shown in Fig. 4(b) were revealed in the 50±
60 dB peak amplitude distribution with a much wider spread
of 1±10,000 event count values. These plots suggest that the
genuine emissions could be distinguished from the mechan-
ical noise signals of inordinate emission bursts.
Fig. 5 represents the AE activity collected on all sensors
during the second load stage of 105±110% maximum oper-
ating load range. The emissions of the second load stage
were high activity compared with those of the first load
stage shown in Fig. 3. In the AE activity plot of Fig. 5(a),
any apparent variations of both cumulative events and peak
amplitudes dependent on load were not observed through
the second load test. These results suggest that the AE
activity was not characteristic of active defect growth