PSI - Issue 5

Jürgen Bär et al. / Procedia Structural Integrity 5 (2017) 793–800 Jürgen Bär et al. / Structural Integrity Procedia 00 (2017) 000 – 000

794 2

da/dN crack propagation rate U actual potential N failure total lifetime N initiation crack initiation lifetime U 0 potential without crack W specimen width

1. Introduction

The lifetime of cyclically loaded components is generally divided into two stages: crack initiation and crack propagation that begins at a technical crack length. Following this division, the crack initiation lifetime is about 70 80 % of the total lifetime, depending on the defined technical crack length. However, investigations performed on technical alloys have shown that cracks are initiated after only a few cycles and consequently, the major part of the cyclic lifetime is covered by the propagation of short fatigue cracks (Polák 2003; Bär 2014). Hence, the propagation of short cracks has to be investigated in detail to obtain realistic lifetime predictions. In this investigation, the short crack propagation at notches is studied on a 7475-T761 aluminum alloy by means of a high-resolution DC potential drop method improved by a special filter algorithm and optical measurements in a SEM. Moreover, the experimental results are compared to FEM calculations to determine the interdependencies of the short crack propagation in different directions. The experiments were carried out on single edge notched specimens of EN AW 7475-T761 clad sheet material. Notches with a depth of 1, 2 and 3 mm have been machined in the specimen with a size of 80 mm x 12 mm x 2.85 mm. The fatigue experiments were performed with constant stress amplitudes under fully reversed loading conditions (R = -1). Fixed grips were used to minimize bending forces. For the potential drop measurement two pins were mounted in the specimen symmetrical to the notch with a distance of 4 mm. The potential was measured with an amplifier of the control electronics. Bär and Volpp (Bär 2001) provide detailed information on the testing equipment. 2. Experimental Details The DC-potential-drop method is widely used to determine the crack length in fatigue crack propagation experiments. Recent developments of this measuring technique have enhanced the resolution of the method as well as the sampling rate. Because of the continuing improvement of the sampling rate, the maximum value within a load cycle, as requested in the ASTM E647 (ASTM 2013), can be determined even at high loading frequencies. However, obtaining potential data for each cycle causes some problems due to scatter of the measured values. The application of the secant method for determining the crack propagation rate according to the ASTM E647 results in high scatter and even negative values are calculated. This indicates that this method is not suitable for determining the crack propagation rate in continuously recorded data. Moreover, investigations undertaken by Donelly and Nelson (Donelly 2002) showed that for low crack propagation rates the uncertainty of the measurement is dominating and not the crack propagation. The second method recommended by the ASTM is an incremental polynomial method. Ostergaard (Ostergaard 1979) has shown that this method in some cases is even less reliable than the secant method. The idea of reducing the number of data points to suppress the scatter could be performed by using a crack length or a load cycle interval criterion. Investigations by Benedictus de Vries et al. (Benedictus-de Vries 2004) have shown that a crack length criterion leads inevitably to mathematical errors which cause a shift and a tilt of the resulting a-N curves. This evaluation of how to handle potential drop data highlights that potential drop data has to be smoothed in an appropriate manner to obtain reliable rack propagation rates. 3. Potential drop measurement