Ultrasonic detectability of potentially closed cracks from cold-worked holes under loaded conditions

Many non-destructive inspections for cracks rely on detecting ultrasound scattered or reflected from the crack. Recently, several methods for second-layer ultrasonic crack detection have been developed, sometimes using automated analysis methods to interpret the complex ultrasonic responses from multiple layers. In the validation of these techniques, the assumption is often made that fatigue cracks will reflect as much ultrasound as an artificially introduced EDM notch or saw cut. However, this is not always the case for closed cracks, which can have a very low ultrasonic reflectivity. Of particular concern is the case of cold-worked holes, where a residual compressive stress field exists, extending to approximately one radius from the fastener hole. In addition, if a crack has grown under fatigue loads that include high peak stresses, there can be a closed region in the plastic zone at the crack tip. To complicate matters further, interference-fit fasteners apply an additional tensile stress, which may or may not counteract the cold-worked stress field near to the fastener, depending on the amount of interference.

This problem was encountered whilst validating a two-layer ultrasonic inspection technique on real cracks in a real fatigue-test wing containing cold-worked holes. Inspections were performed at four different remote load levels by stress-jacking the wing. Whilst the ultrasonic indications from some of the cracks grew with increasing remote load, others did not. Some ultrasonic indications grew both towards the fastener and away from it, suggesting that both crack-tip closure and cold-worked crack-root closure were occurring. A possible reason for some cracks not growing is that they were already being held open by interference-fit fasteners.

The aim of this project was to investigate and understand the effects of cold working, interference-fit fasteners and crack-tip closure on the ultrasonic detectability and sizing of cracks. In particular, information was required about the potential improvement in defect detectability by applying a remote tensile load to the structure to overcome the residual compressive stresses. Results will be reported of an experiment designed to grow fatigue cracks to a range of lengths at cold-worked holes in a single specimen, with interference-fit fasteners installed. These cracks could then be inspected ultrasonically at various remote tensile-load levels and the results compared with EDM notches in a similar specimen. Finally, the interference-fit fasteners could be removed and the ultrasonic inspections at the various loads repeated. Initial results have confirmed that ultrasonic indication size and amplitude both increase with remote tensile load but further results will be required to determine the dependence of defect detectability on remote load.