In this study, we focused at the development and verification of a robust framework for surface crack detection in steel pipes using measured vibration responses; with the presence of multiple progressive damage occurring in different locations within the structure. Feature selection, dimensionality reduction, and multi-class support vector machine were established for this purpose.<p> </p>Nine damage cases, at different locations, orientations and length, were introduced into the pipe structure. The pipe was impacted 300 times using an impact hammer, after each damage case, the vibration data were collected using 3 PZT wafers which were installed on the outer surface of the pipe. At first, damage sensitive features were extracted using the frequency response function approach followed by recursive feature elimination for dimensionality reduction. Then, a multi-class support vector machine learning algorithm was employed to train the data and generate a statistical model. Once the model is established, decision values and distances from the hyper-plane were generated for the new collected data using the trained model. This process was repeated on the data collected from each sensor. Overall, using a single sensor for training and testing led to a very high accuracy reaching 98% in the assessment of the 9 damage cases used in this study.
Investigated is the ability of ultrasonic guided waves to detect flaws and assess the quality of friction stir welds (FSW). AZ31B magnesium plates were friction stir welded. While process parameters of spindle speed and tool feed were fixed, shoulder penetration depth was varied resulting in welds of varying quality.<p> </p>Ultrasonic waves were excited at different frequencies using piezoelectric wafers and the fundamental symmetric (S0) mode was selected to detect the flaws resulting from the welding process. The front of the first transmitted wave signal was used to capture the S0 mode. A damage index (DI) measure was defined based on the amplitude attenuation after wave interaction with the welded zone. Computed Tomography (CT) scanning was employed as a nondestructive testing (NDT) technique to assess the actual weld quality. Derived DI values were plotted against CT-derived flaw volume resulting in a perfectly linear fit. The proposed approach showed high sensitivity of the S0 mode to internal flaws within the weld. As such, this methodology bears great potential as a future predictive method for the evaluation of FSW weld quality.