Issue 48

J. Bär et alii, Frattura ed Integrità Strutturale, 48 (2019) 563-570; DOI: 10.3221/IGF-ESIS.48.54 565 The experiments were undertaken with constant stress intensities K max = const. and  K = const.. The crack initiation and the first part of the crack propagation experiment were performed with a constant force amplitude until the desired stress intensity, due to the increasing crack length, was achieved. Henceforward, the experiments were performed with constant stress intensities, K max = 5; 7.5 and 10 MPa m , until a maximum crack length of 8 mm, respectively. For each stress intensity level, five specimens have been tested. Lock-In Thermography The fatigue crack propagation experiments were accompanied by thermographic measurements performed with a CEDIP Titanium HD 560 camera. During the crack propagation experiments, the camera automatically recorded sequences of 990 frames with a framerate of 99 Hz in defined intervals of 1 min. Each frame consists of 640 x 512 pixel with a spatial resolution of 29 µm/pixel. For the Lock-In analysis the load, delivered from the control electronics of the testing machine in form of an analogue signal, was recorded by the camera in form of a 14-bit value. To achieve a high and equal emissivity, the surface of the specimens was grinded and coated with a layer of black paint with a thickness of about 20 µm. The Lock-In evaluation of the thermographic measurements was undertaken with a self-developed MatLab program, including a motion compensation algorithm to compensate the rigid body movement of the specimen during a sequence [11]. The Lock-In evaluation was performed for each pixel of the recorded frames using a Discrete Fourier Transformation according to Eq.(3). Figure 1 : Resulting amplitude images of the DFT according to Eq. (3): The E-Mode (a) is coupled with the loading frequency, the D-Mode (b), the D1-Mode (c) and the D2-Mode represent the first, second and third harmonic. For each analyzed sequence of 990 frames, the evaluation provides an image for the average temperature T m , as well as, amplitude and phase images for the E-, D-, D1- and D2-Mode, respectively. The DFT evaluation reduces the noise down to a value of 1.6 mK. An example for the resulting images for a specimen loaded with K max = 10 MPa m and a crack length of a = 7.1 mm is shown in Fig. 1. To compare the results of the thermographic measurements obtained in different experiments, the results were quantified by determining the arithmetic mean of all amplitude values and the maximum value within defined areas. For a detailed examination, a separate evaluation was undertaken for the crack tip and the crack flanks (Fig. 2). At the crack tip an area of 73 x 73 pixels (2.05 x 2.05 mm) was located axially symmetric to the notch root. This area was shifted automatically in the