The fracture dynamics were studied by the S0-packet waveform matching. Detail of this method can be found elsewhere [6]. The overall transfer function, including the sensor, medium and AE monitoring system, was determined by the method shown in Fig. 6. Lamb-wave AE signals were exited by a compression-mode PZT element, glued to a specimen and stainless steel block by epoxy, and were monitored by PICO sensor. Here, the out-of-plane displacement of PZT element was monitored by a laser interferometer and converted to time transient of crack volume. The overall transfer function for the Lamb-wave due to Mode-I crack opening was obtained by the time-domain Gauss-Zeidel deconvolution of the detected waves by the PICO sensor. Crack opening was represented by a sinusoidal rump function with rise time, ∆Tr, and crack volume, ∆V. They were determined by matching the simulated waveform of the S0-packet to the detected one. SUS block PZT Element ModeⅠ fracture Small AE sensor ch.1 4 Propagation distance (3t×4w) Pulse generator Toughened Si3N4 3 Laser interferometer Figure 6: Experimental setup to determine the overall transfer function. 20 21 22 23 0 1 2 600 700 800 Time ks Crack volume ∆V 10–16 m3 Maximum stress σ max MPa fracture Maximum stress σ max Final 1 2 (b) 107 108 109 110 0 1 2 600 700 800 Time ks Crack volume ∆V 10–16 m3 Stress σ MPa fracture Stress σ Final 1 2 3 4 (a) Figure 7: AE source volumes as a function of applied stresses for the quasi-static(a) and the cyclic(b) loading. Figure 7 shows ∆V estimated for quasi-static (a) and cyclic (b) loadings with stress values as a function of testing time. The numbers in the figure indicate event count. The crack volume produced during the cyclic loading was estimated to be six times smaller than that produced during static loading at lower stress level.
RkJQdWJsaXNoZXIy MjM0NDE=