Practical NDE techniques by ultrasonics are based on A, B and C scan images processed and filtered in a variety of ways, from which the characterization of the defects are quantitatively extracted by heuristic methods or even by visual interpretation of the image. Apart of the arbitrariness of the choice of this threshold, a great amount of information contained in the image and therefore in the measured signal is being lost. This rich data may be crucial to combat the noise that hides difficult defects. This paper presents an effort to integrate all the information recorded in the measurements in a generalized processing or inversion scheme. QNDE was originated as an application of the fast developing numerical methods to so-called inverse problems. A number of works have been developed for idealized probes with emphasis on the numerical techniques, but using a vaguely developed link to the measurement and search procedure. In this paper, such a numerical procedure is extended and deployed to an experimental case. The principle to be used is the measurement and inversion of frequency-domain information instead of classical time-domain delays or vibration eigenmodes or eigenvalues, together with the use of a reduced set of output data understood as a regularization technique to drastically overcome noise problems. A deconvolution scheme from a sane specimen overrides uncertainties about the input signal and other coherent noise. The main strength of this approach is that it is not necessary to visually be able to identify the portion of the signal that contains the information about the flaw, which may be hidden under many complicated patterns or other waves, The approach is used for the experimental case of an aluminum specimen with a defect in the form of a side drilled, hole. The ultrasonic measurements are using a simulated array of transmitters as well as receivers. A wedge transducer has been developed and manufactured for this technique to achieve point contact. Neither of them are placed right above; the defect, and the magnitude of surface waves hides the echo from the defect. Despite these facts, this inversion technique successfully finds the depth and size of the defect.
展开▼
