| Popis: |
With the industrial use of friction stir welding (FSW) growing steadily, nondestructive testing (NDT) methods have to be developed allowing the reliable detection of existent nonconformities in the weld during a post-process inspection. Moreover, there is also a strong need for better process monitoring and control by in-process NDT methods. In case of FSW (the risk of) a defect formation have to be detected during welding in order to optimize the process in regards to weld quality. At IZFP we are engaged in both aspects of FSW NDT. In the present contribution we will describe the application of well-established but also novel NDT methods for defect detection in FSW. Besides "standard" weld irregularities like lack of fusion, lack of penetration, porosity, inclusions, cracking and undercut our main focus was on defects, affecting the fatigue limit and other characteristics of dynamic loading. Oxide and hydroxide fragments stirred into the aluminium weld as fine lines are examples for such fatigue-affecting defects. Multi-frequency eddy-current, ultrasound immersion, dry ultrasound testing with EMAT (electro-magnetic acoustic transducer), ultrasound backscattering, x-ray radiography and computer tomography are NDT methods, which have been investigated in regards to their ability to detect such defects. Concerning FSW monitoring and control, it could be shown, that the distribution of oxide lines could be correlated with certain welding process parameters. Consequently, an approach was suggested to prevent the conglomeration of oxide fragments by several measures. Furthermore it was demonstrated, that a special analysis of monitored welding forces can be used in order to predict the formation of different weld irregularities, like oxide fragments and elongated cavities (wormholes). Based on this approach a control algorithm was developed, which allows the real-time adaptation of the manipulated variables of the FSW process in order to prevent the formation of these defects. |
