PSI - Issue 68

Andreas J. Brunner et al. / Procedia Structural Integrity 68 (2025) 1266–1272 Brunner et al. / Structural Integrity Procedia 00 (2025) 000–000

1271

6

half of the ply thickness of the unidirectional prepreg (about 780 micrometer, from 4.42 – 4.92 mm specimen thickness divided by six plies). It could be hypothesized that meso-scale morphology may play a role in determining defect size distributions. Confirmation, of course, will require further testing of laminates with a sufficiently different range of ply thicknesses.

Table 5. Corrected estimate of average delamination area and diameter per AE signal for GF-EP2 under different loading modes (see text for types of correction applied)

Loading Mode

Remarks

Fracture Area ( mm 2 )

Number of AE signals from process zone ( - ) 19’769±2’105 7’337±3’084

Average defect area ( µ m 2 ) per AE signal

Average defect diameter ( µ m ) square-root of area

Mode I tensile opening

2’129±127 1’300±107

108’702±11’308 226’367±120’642

329±17 460±120

Fixed Ratio Mixed Mode I/II (4:3 ratio of I:II) Mixed Mode Bending I/II (tested at 4:3 ratio of I:II)

768±22

46’505±14’351

17’889±4’632

133±18

One sensor only, hence area correction only

Mode II in-plane shear

1’142±76

15’493±9’445

152’548±141’047

352±170 380±133

Average ± Standard deviation

-

-

Without Mixed Mode Bending

3.4. Assessing AE signal amplitudes and defect size distributions For micromechanical modelling, defect size distributions, not only averages, are essential. Such defect size distributions result from a sensitivity factor that correlates average AE signal amplitudes in mV with average defect diameter in µ m. This approach is inspired by Baensch et al. (2015) who correlated AE data with imaging defect sizes in wood materials by synchrotron X-ray micro-computed tomography. Analogous to that, delaminated fracture surface area can be correlated with cumulated AE signal amplitudes. The "sensitivity factor" from this correlation then yields an estimate for defect size for given amplitude values. For GF-EP2, Gfrerrer et al. (2024) determined defect size distributions as well as estimates of minimum detectable defect sizes. If AE signal amplitudes are measured in dB AE instead of in mV and the number of AE signals is chosen to represent signal sources located in the fracture process zone, these distributions are roughly symmetrical around the average defect diameters. For AE signal amplitudes in (linear) mV-scale, the distributions are asymmetrical, and result in lower defect diameters than from the sensitivity factor from the (logarithmic) dB AE -scale. 4. Summary and outlook The combination of quasi-static fracture tests with AE monitoring allows for roughly estimating average values of microscopic defect sizes that result in macroscopic delamination propagation under different loading modes. Even if the input data are corrected for effective number of AE signals contributing to delamination formation and for effective resulting fracture surface area, the average order of magnitude for GF-EP2 amounts to between 300 and 450 micrometer in diameter. The AE signal amplitude distributions correlate with defect size distributions. Variations in the speed of delamination propagation essentially yield increasing and decreasing number of AE signals per unit time, respectively rather than changes in the range of defect diameters. The chosen AE signal acquisition threshold yields an estimate of the smallest detectable damage. Rough estimates of the relevant time-scale, i.e., the duration of the damage events, from AE source modelling, including signal propagation and transfer function of the measurement chain, yield source rise-times between a few ten of nanoseconds and a few microseconds, see Sause et al. (2010) for details. Hence, time-steps of a few nanoseconds (5-10 ns) may be required for sufficiently accurate modelling.