13th International Conference on Fracture June 16–21, 2013, Beijing, China -7- groups are selected to give the detailed results in Fig. 5. Figure 5. Detailed localization results of the point 4, point 10 and point 17. The solid points represent the results of the modified method and the empty points of the classical method. 4.3. Onset time determination Source localization is based on the onset times or arrival times of the direct body waves. Theoretically, onset times of the direct body waves, P-wave and S-wave both can be used for source localization. However, only first wave onset times (P-times) are usually measurable, since multiple side reflections, structural noise and sensor response will interfere the later phases. The true onset time of a crack AE event could be described as the moment when the first energy of a particular signal phase reaches the sensor positions [17]. In signal analysis, the onset time is usually picked as the point where the first difference between the signal and the noise takes place [18]. Determination of the onset time of AE signals is crucial for the whole localization process and is the major premise to affect the localization accuracy [6]. Manually picking is the preferred choice if there are only few events available. It is still necessary to have reliable automatic picking tools, because human analysts cannot manage the vast amount of data recorded in the monitoring. In concrete monitoring, many algorithms have been proposed for automatic detection of onset times and the detailed description can be found in [19]. Two mostly used methods, STA/LTA (STA Short Term Average, LTA Long Term Average) picker by [20] and AIC-picker [6, 21] are employed and investigated for the AE signals in masonry structures. The picking accuracy of the STA/LTA-picker and AIC-picker was quantified by a comparison with picks that were set manually. For 1122 signals of AE events recorded at the pencil-lead break tests in Sections 3, the differences t between the automatic and the manual picks are examined. Fig. 6 shows the results for the STA/LTA-picker and for the AIC-picker, compared to manually picked onset times.
RkJQdWJsaXNoZXIy MjM0NDE=