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

1268

3

2. Materials and methods A first set of GF-epoxy (labelled GF-EP1) consisted of woven GF layers type 92112, EC 9-136 / EC 9-136 and areal weight of 200 g/m 2 (Suter Kunststoffe AG, Jegenstorf, Switzerland) in a matrix from CY225 resin and hardener HY925 (Huntsman, Basel, Switzerland). GFRP laminates were produced by wet lay-up of 22 plies by vacuum bagging with a FEP release film (type Diatex Halar P3, 12.5 μm thickness; Suter Kunststoffe AG) at mid-plane. Curing was under pressure for 2 hours at 140°C and post-curing in an oven for 10 hours at 140°C, resulting in 3.86-3.92 mm thick specimens. A second set of GF-epoxy (labelled GF-EP2) comprising six UD layers was manufactured in an out-of autoclave vacuum bagging process from HexPly® M79-LT/25%/UD1200/G prepreg with a non-adhesive starter film embedded at the mid-plane on one side of the plate. Curing of the resin was by heating from room temperature to 80 °C within 80 min, keeping the temperature for 6 h followed by cooling down to room temperature within 25 min. GF epoxy specimens were between 4.42 and 4.92 mm thick with a fiber volume fraction of approximately 60 %. Specimens were roughly 20 mm wide. CF-thermoplastic (AS4/PEEK) specimens were provided within a joint round robin between ASTM, ESIS and JIS, see NASA-TM-104222 (1992). AE monitoring for data in Table 1 was performed with a SPARTAN equipment (Physical Acoustics Corp.) with one R15 type sensor, for data in Table 2 with an AMSY 6 equipment (Vallen Systeme GmbH) with two SE150-M type sensors, and for data in Table 3 with a MISTRAS equipment (Physical Acoustics Corp.) with one or two WD type sensors. 3. Results and discussion 3.1. Simple estimates without corrections In order to investigate the applicability of the "simple" estimate by Brunner (2016) for average defect size from dividing the projection of the fracture surface by all recorded AE signals (Table 1), the procedure was applied to GF EP1 (Table 2). The estimated average defect diameter for GF-EP1 is slightly higher than that for AS4/PEEK (Table 1). Data for GF-EP1 evaluated for a second sensor mounted on the load-block yield higher numbers of AE signals and hence somewhat lower defect diameters (-10% to -15%), but still within one standard deviation. The number of recorded AE signals depends on the material properties, e.g., signal attenuation, the AE equipment, sensor type and sensitivity, on sensor position and coupling as well as on the AE signal acquisition settings.

Table 1. Simple estimate of average delamination area and diameter per AE signal for an AS4/PEEK laminate under Mode I tensile opening load Fracture Area ( mm 2 ) Total number of AE signals ( - ) Average defect area ( µ m 2 ) per AE signal Remarks

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

Specimen 1 Specimen 2 Specimen 3

538 459 432 476

35'620 32'790 51'970 40'127 10'354

15'104 13'998

123 118

AE sensor near specimen end AE sensor near specimen end AE sensor near specimen end

8'312

91

Average

12'472

111

- -

Standard deviation

55

3'644

17

Table 3 presents "simple" average estimates for GF-EP2 tested under different quasi-static fracture loading modes. Comparing Mode I for GF-EP1 and GF-EP2, the estimates agree within one standard deviation, with GF-EP2 yielding slightly higher values. For GF-EP1, the reinforcement consists of woven GF, and for GF-EP2 of unidirectional GF plies. Hence, the difference due to reinforcement type is possibly small. The comparison between loading modes (Table 3) indicates differences between 100 and 150 micrometer for "simple" average estimates, but when considering the standard deviations, the differences are hardly significant. The largest nominal difference is observed between the two types of Mixed Mode I/II tests, both performed at the same mode mixity ratio (4:3 for I:II). Different test rig design may affect the AE data recording, but the total delaminated area also differs between the two test rigs. Under