부산대학교 도서관

An automated data acquisition system has been developed that can be used as an independent watchdog to monitor the loads applied during a fatigue test. An algorithm to perform peak-picking based on a quadratic least squares approach has been shown to yield peak/valley magnitudes to within 1%. A damage parameter is described that assesses the impact of loading error on the resulting fatigue crack growth behavior. The peak-picking system has been validated for both constant amplitude as well as variable amplitude, spectrum loading. Overall test system performance is shown to vary significantly during the spectrum testing for the three different types of controllers: an analog controller with a computer control fatigue test system and two different types of digital controllers. This testing and analysis clearly demonstrates the importance of quantifying range error (this is not the same as end-level error), number of occurrences for each range, and overall damage content. Finally, the importance of standardizing performance metrics for spectrum testing methods is presented along with some potential recommendations for what these metrics could be and how they might be developed.