PSI - Issue 28

F.W. Panella et al. / Procedia Structural Integrity 28 (2020) 1709–1718 Author name / Structural Integrity Procedia 00 (2019) 000–000

1717

9

as a function of thickness distance on the A-scan display (Garnier et al. (2011)), or more simply analyzing signal peak presence and location. Amplitude data are recorded at regular intervals each of 10000 cycles for each B-type and W-type specimens and resulting data are analyzed as function of probe position on specific sample ROIs; two central ROI are selected for both test cases, one between the bottom support spans for CFRP bar and the other one around the hole section, supposed as failure regions by DIC results from static preliminary tests. First results are quite satisfactory with significant deviations from initial trends, as shown in Fig. 12b, where a small damage occur beside load supports are observed at 65% of fatigue life for B11 sample. UT A-scan signal amplitudes of real detected defects were inspected using PA UT technology are directly compared with those of artificial defects in similar material section at equivalent depth employed for DAC curves, as in Fig. 11a and 11b. An internal delamination could be observed at a depth of 2.98 mm in Fig. 11b, where an approximate section of 7 × 1.2 mm could be estimated. UT results may depend on contact condition, type of coupling, human error, calibration or software data settings; therefore, the repeatability tests are done on specific ROIs of W1 sample, on the central hole zone of wrinkle samples in different days to demonstrate method efficiency and contained data variability in different conditions, as in Fig. 13a. Amplitude data are recorded at regular intervals and resulting data are analyzed as function of probe position on selected ROI of W1 sample on central section with hole zone along load direction. First results are quite satisfactory with significant deviations of 20% and 60% at 80% and 90% fatigue life, as shown in Fig. 13b where each UT scans are obtained at different fatigue life of W1 sample on central section with hole zone along load direction.

a

b

Fig. 11. (a) A scan and S scan at for 0% vs (b) 65% fatigue life of bar B11 sample.

Conclusions Fatigue experimental results highlights general data variation in damage modes and location due to test variability, but even that different fatigue failure zones of specimens are evaluated, good coherence between NDT damage levels and stiffness life is observed. Fatigue experimental results indicated fatigue life of CFRP elements is determined and anticipated by continuous and early damage nucleation and structural specimen state, well before final rupture. The suitable implementation of experimental fatigue tests and NDT controls seem indicate appropriate strategy for damage prevention and evaluating fatigue effects on specific zones of CFRP materials. In the same time, uncalibrated signal has been analyzed in the time domain for extracting information related to the temperature increase, thermoelastic and dissipative sources and phase shift between imposed stress and strain, according to damage accumulation. Since results show good agreement of damage accumulation as monitored experimentally with obtained fatigue experimental curves, proposed procedure could represent useful tool to investigate on fatigue behavior of critical components, even of complex shape, subjected to variable loading, through damage nucleation detection well before failure.