STP722

    Automatic Detection, Classification, and Sizing of Steam-Generator Tubing Defects by Digital Signal Processing

    Published: Jan 1981


      Format Pages Price  
    PDF Version (148K) 10 $25   ADD TO CART
    Complete Source PDF (10M) 10 $55   ADD TO CART


    Abstract

    An eddy-current analysis system is being developed for the inspection of steam-generator tubing that provides automatic interpretation of eddy-current signals. The microprocessor-based system realizes real-time analysis capability through a programmed probe controller, on-line data conversion, and high-speed digital signal processing. In a fully automatic mode, a wide variety of anomalies can be detected by appropriate enhancement of their characteristic wave shapes. Each detected signal is transformed and parameterized to yield a set of representative features that are used to classify the type of anomaly. An additional feature set is then extracted for estimating the depth of the discontinuity. Classification and sizing capabilities are provided by adaptive learning networks that were trained on a data base of simulated-defect waveforms. Machined defects were utilized in training to simulate as much as possible the most prevalent problems of stress-corrosion cracking, pitting, wastage (wall thinning), and denting of steam-generator tubing. The data base includes combinations of these defect types as well. The range of depth estimation is from 10 to 100 percent of the tube-wall thickness. The system has been designed to interface with either single-frequency or multifrequency devices.

    Keywords:

    eddy current analysis, automatic interpretation, on-line data conversion, digital signal processing, parameterization, feature extraction, classification, adaptive learning networks, stress corrosion cracking, pitting, wastage, phase plane pattern, nondestructive evaluation, Inconel 600 tubing, program controlled devices, spectral measurements, incremental encoder


    Author Information:

    Brown, CL
    Scientist/engineer, engineering aide, engineering aide, and vice-president, Adaptronics, Inc., McLean, Va.

    Defibaugh, DC
    Scientist/engineer, engineering aide, engineering aide, and vice-president, Adaptronics, Inc., McLean, Va.

    Morgan, EB
    Scientist/engineer, engineering aide, engineering aide, and vice-president, Adaptronics, Inc., McLean, Va.

    Mucciardi, AN
    Scientist/engineer, engineering aide, engineering aide, and vice-president, Adaptronics, Inc., McLean, Va.


    Paper ID: STP27605S

    Committee/Subcommittee: E07.11

    DOI: 10.1520/STP27605S


    CrossRef ASTM International is a member of CrossRef.