ABSTRACT
Ultrasonic nondestructive testing of steel forgings aims at the detection and classification of material inhomogeneities to ensure the components fitness for use. Due to the high price and safety critical nature of large forgings for turbomachinery, there is great interest in the application of imaging algorithms to inspection data. However, small flaw indications that cannot be sufficiently resolved have to be characterized using amplitude-based quantification. One such method is the distance gain size method, which converts the maximum echo amplitudes into the diameters of penny-shaped equivalent size reflectors. The approach presented in this contribution combines the synthetic aperture focusing technique (SAFT) with an iterative inversion scheme to locate and quantify small flaws in a more reliable way. Ultrasonic inspection data obtained in a pulse-echo configuration are reconstructed by means of an Synthetic Focusing Technique (SAFT). From the reconstructed data, the amount and approximate location of small flaws are extracted. These predetermined positions, along with the constrained defect model of a penny-shaped crack, provide the initial parametrization for an elastodynamic simulation based on the Kirchhoff approximation. The identification of the optimal parameter set is achieved through an iteratively regularized Gauss-Newton method. By testing the characterization method on a series of flat-bottom holes under laboratory conditions, we demonstrate that the procedure is applicable over a wide range of defect sizes. To show suitability for large forging inspection, we additionally evaluate the inspection data of a large generator shaft forging of 0.6-m diameter.
