PSI - Issue 38

S. Häusler et al. / Procedia Structural Integrity 38 (2022) 230–237

231

2

S. Häusler et al. / Structural Integrity Procedia 00 (2021) 000 – 000

1. Introduction Despite the advanced development of calculating the fatigue strength of rotor blades for wind turbines, the damage state during operation remains unknown. To still ensure a safe lifetime for 20 to 25 years the blades are often overdesigned. In contradiction, the turbine has to be removed after its design lifetime even though structural reserves for a prolonged operation exist. To determine the real time damage state during operation, in the DynaWind² project a progressive damage model based on experimental data will be developed. For providing a sufficient data basis for the numerical calculations, the mechanical properties of glass fibre reinforced plastics (GFRP) under fatigue loading must be characterised. Hereby the stiffness and strength degradation are of special interest since stiffness degradation from 10 up to 45% can occur, Nijssen (2006), and consequently have a high impact on the materials behaviour under operational load conditions. Various approaches have been made to characterise the degradation properties of composite materials. While the load cell in the testing machine measures the applied forces reliably, the recording of valid strains in the area of examination is quite difficult. Using the machines displacement data as in Cadavid et al. (2017) and Roundi et al. (2019), the material ’ s fatigue behaviour within the region of interest is usually not provided. The measurement contains the influences of the clamping area like tab debonding or damage transition from the centre into the tab region as well as the machine’s stiffness and friction effects. A relative analysis can be utilised, but evaluating the absolute material degradation must be treated with caution since the previously mentioned influences have to be considered. Another approach is the measurement with strain gauges or extensometers directly applied on the surface of the specimens, Zangenberg et al. (2014) and Wedel-Heinen et al. (2006). In general, these approaches involve the risk that the gauges detach during high cycle fatigue tests, as experienced by Nijssen (2006). In addition, conventional strain gauges offer only a localised measurement, which is significantly influenced by the relative position of the damage with respect to the gauge. In a worst-case scenario, the strain gauge measurement of a fatigue test can be lost due to a crack or delamination directly below the gauge area. In this paper an approach is presented to measure the strains during fatigue tests using the contactless digital image correlation method. In order to track the damage evolution and crack propagation during the experiment, backlight images are taken by an additional camera. Therefore, specimens with 90° and ±45° fibre angle have been tested in high cycle fatigue with stress ratios ( R ) of 0.1 and -1. With the proposed experimental setup, S-N curves and stiffness degradation curves are obtained. For the analysis, a previously developed evaluation method for strain measurement on GFRP, Häusler et al. (2021), has been applied.

Nomenclature GFRP glass fibre reinforced plastic DIC digital image correlation R stress ratio Θ

fibre angle compared to the long side of the specimen

2. Experimental setup To determine the fatigue performance of GFRP, experiments have been conducted with the parameters shown in Table 1. In accordance to the progressive damage model’s requirements , the experiments were performed using the pearl string method DIN (2016). Therefore, the stress amplitudes were chosen to reach 10 4 , 10 5 and 10 6 load cycles until final failure, respectively. In each of the first three sets, displayed in Table 1, at least three specimens were tested per load cycle level. They were tested in an Instron 100 kN 8801 servo-hydraulic testing machine. Before the constant amplitude fatigue experiment started, t he Young’s modulus was measured quasi-statically in the tensile range of the respective stress amplitude. The strains were measured with the DIC method, which requires a speckle