PSI - Issue 17
Meike Funk et al. / Procedia Structural Integrity 17 (2019) 183–189 Author name / Structural Integrity Procedia 00 (2019) 000 – 000
185
3
Fig. 1. (a) Sketch of a SEN specimen with two bonded wire pairs for the DCPD. (b) Photo of a SEN specimen with welded wires.
1.2. DCPD
One key topic is the crack length measurement during a fatigue test which can be realized with a DCPD-Method. Changes in the potential drop are directly related to crack length (ASTM E647), as described in equation 1 (Johnson (1965)). For the calculation, the potential drop U is normalized to the potential of the crack-free specimen U 0 .
w y
cosh
0
w a
arccos 2
2
=
(1)
w y
cosh
0
0 U U
2
cosh
arccosh
a
cos
k
w
2
A crack growth should be detectable by a significance of 5% (ASTM E647), which corresponds to an increase of 1.96 times the standard deviation. The used set-up has an initial signal of U 0 = 0.14419 ± 0.0001 mV, with y 0 = 2 mm and I = 45 A. Thus, starting from the uncracked specimen, a first crack growth can be detected when the potential rises by ΔU = 0.0001 mV which corresponds to a crack length of about 4 µm. With increasing crack length a , the calibration curve ( a / w over U / U 0 ) flattens out, resulting in an increase in the sensitivity of the DCPD and a decrease in the influence of noise. The resolution of the DCPD becomes better with increasing crack length a up to less than Δa = 1 µm (Fig. 2b). The recorded potential drop is strongly dependent on the shape of the notch. Blunt notches give an error in crack length measurement due to a change in the electrical potential lines (ASTM E647; Allery and Birkbeck (1972); Johnson (1965)). Thus, there is a difference between the real bonding distance y 0,real and the bonding distance adjusted to the Jonson formula y 0,Johnson , this can be done by finite elements calculations or by calibration of the system with defined crack length.
Made with FlippingBook Digital Publishing Software