PSI - Issue 16

Roman Chepil et al. / Procedia Structural Integrity 16 (2019) 211–217 Roman Chepil, Olena Stankevych, Orest Ostash, Bohdan Klym / Structural Integrity Procedia 00 (2019) 000 – 000

214

4

initiation period N i . Then, the dependence (4) may be written as

,

(5)

N N 

f

i a d 

*

i

and lifetime N f can be predicted using the base notch fatigue curve for macrocrack initiation stage (Ostash et al. (2017)). This curve (Fig. 2) is built from the test results of standard compact specimens with a different notch radius  in coordinates * y  vs. N i . The relationship (5) can be the basis for period N i prediction in notched components of various shapes with high accuracy (Ostash et al. (2017)).

Fig. 2. Base notch fatigue curve for CT- specimens made of D16pchAT aluminum alloy ( d * = 250  m).

4. Fatigue fracture monitoring at the macrocrack initiation stage

To validate the predicted durability of structural elements the accumulation of damage during their operation should be monitored. This is especially true for cases where an abrupt fracture can occur after the initiation of a fatigue macrocrack of the length a i = d * , when d * = 0.1  0.3 mm (Ostash et al. (2001)). It is often difficult to detect such cracks by traditional non-destructive flaw detection. Therefore, a regularity of the damage accumulation (fatigue microcracks initiation and growth, and macrocrack formation) with the simultaneous analysis of acoustic emission signals (AE) generation were studied. The notched compact tension (CT) specimens (Fig. 2) of width W = 40 mm, the thickness of 4.35 mm and notch root radius  = 1.6 mm made from a rolled sheet of D16pchAT aluminum alloy (type 2024-T3) were investigated. The t est was performed in laboratory air at 20  22  С under uniaxial constant amplitude loading cycle with the frequency 7  8 Hz and stress ratio R = P min / P max = 0.1. An observation of the area near notch root was carried out using a microscope MBS-10. To minimize the effects of signals generated by the test equipment, the specimen was isolated with a dielectric material. The experiment was stopped and the specimen photo was taken at increasing of AE activity. To visualize microcracks and to determine the area of initiated macrocrack into the stress concentrator tip a paint was periodically applied. Two-channel AE detection system SKOP-8M (Nazarchuk and Skalsky (2009)) was used to record the AE activity during specimen testing. The system is completed by a band sensor with an operating range of frequencies between 200 and 600 kHz. AE measurement conditions of pre-amplifier 40 dB, the threshold of discrimination within 30% and sampling rate 4 MHz were adopted. Band-pass filtering of 200 – 600 kHz was performed. The relative error of registration of AE signals amplitude is less ± 10%. The acquired waveforms were then stored on a computer for further analysis and display. For analyses of the local peculiarities of AE signals, the next algorithm was proposed (Stankevych and Skalsky (2016)): (1) the realization of the continuous Wavelet transformation (CWT) for the AE signal (Fig. 3a): (2) the determination of the maximum wavelet coefficient WT max of the single AE event, corresponding frequency f max , bandwidth  f by 0.7 WT max level of the 2D frequency projection (Fig. 3b); (3) the determination of the time interval [ t 1 ; t 2 ] and pulse emission duration  t by 0.5 WT max level of the 2D time projection (Fig. 3c); (4) the determination of the approximation coefficients A , B , C , D for the function WT a ( t ) which