Probability of detection in friction stir welding using nondestructive evaluation techniques

Abstract

When available, the force feedback data from Friction stir welding (FSW) can be very useful for analyzing weld quality. Friction Stir Welding Analysis Tool software (FSWAT), a new process based non-destructive evaluation (NDE) technique developed at the South Dakota School of Mines and Technology, is designed to analyze any specified section of FSW in real time (with a slight computational delay). With this software, a trained operator or inspector can detect where potential flaws may exist. Another powerful NDE technique is ultrasonic phased array (UPA), which is well known for its capability to detect different kinds of friction stir welding indications and defects. The purpose of this study is to compare the defects identified in a round robin investigation using UPA and electromagnetic radiation X-ray inspection with the defects identified by the FSWAT software data analysis program. In addition, actual destructive tests are used to correlate the identified defects with actual defects. A probability of detection (POD) analysis is carried out to evaluate the wormhole-detection performance of the different NDE methods applied currently in industry. By correlating this software with UPA and X-ray inspection, the time and expense associated with 100% inspection of parts could be considerably reduced. The ultimate goal of this research is to support the development of real-time quality control to minimize the cost of inspection.