PSI - Issue 5

Igor Shardakov et al. / Procedia Structural Integrity 5 (2017) 210–216 Igor Shardakov / Structural Integrity Procedia 00 (2017) 000 – 000

6

215

Fig.6. Criterion K versus the distance between the impact point and accelerometer D1

The analysis of the results given in Fig.6 indicates that the dependence of the value K on a distance between the point of impact and the nearest accelerometer D1 is non-monotonic. This non-monotonicity property becomes more pronounced when the pulse time decreases. Using these dependencies, we have found the distance L * between the accelerometer and the point of impact at which the criterion K exhibits the best response to the change in the state of the crack. T he maximum values  К max from Fig.6 correspond to these values of L * . The values of the optimal distance L * calculated for three pulse duration values turned out to be rather close and, for a sequence of pulse duration values T = {0.27, 0.10, 0.05} msec, they are L * = {334, 281, 307} mm, respectively. Figure 7 shows the histograms of the values of the criterion K, which correspond to four states of the crack (values along the axis of abscissa). Each histogram is plotted for specific pulse duration and the relevant value of an optimal distance to the point of impact. A comparison of these histograms demonstrates that the sensitivity of the criterion K to the state of a crack increases as its duration decreases provided that a distance between the point of impact and the accelerometer D1 is taken optimal. The results of the physical experiment have supported the validity of a our version of the shock wave method used to control the condition of the beam. According to the scheme in Fig.1, a beam with a crack in its central section is provided with two accelerators mounted at a distance 70mm between them (symmetrically with respect to the crack). The beam was subjected to a force pulse of 0.27msec at a distance of L=334mm to the nearest accelerator. After the experimental acceleration records from the piezosensors D1 and D2 were processed, the values of the criterion K were established for all the four states of the crack. These experimental values are presented as the shaded bars on the left histogram ( T = 0.27msec). It has been found that the agreement is fairly good between simulations and experimental data.