Calibration and flaw detection method for ultrasonic inspection of acoustically noisy materials

• US 5,445,029 A
• Filed: 11/08/1993
• Issued: 08/29/1995
• Est. Priority Date: 11/08/1993
• Status: Expired due to Term

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Falsetti, Robert V., Dixon, Elizabeth L.
• Primary Examiner(s): Williams, Hezron E.
• Assistant Examiner(s): Oda, Christine K.
• Application Number:
  US08/149,208
• Time in Patent Office: 659 Days
• Field of Search: 364/507, 73/609, 73/602, 73/596, 73/627, 73/575
• US Class Current: 73/609
• CPC Class Codes: G01N 2291/011 Velocity or travel time
• CPC Class Codes: G01N 2291/015 Attenuation, scattering
• CPC Class Codes: G01N 2291/048 Transmission, i.e. analysed material between transmitter and receiver
• CPC Class Codes: G01N 2291/056 Angular incidence, angular propagation
• CPC Class Codes: G01N 2291/102 one emitter, one receiver
• CPC Class Codes: G01N 2291/2632 flat
• CPC Class Codes: G01N 29/2487 Directing probes, e.g. angle probes directing arrangements G01N29/221
• CPC Class Codes: G01N 29/265 by moving the sensor relative to a stationary material
• CPC Class Codes: G01N 29/30 Arrangements for calibrating or comparing, e.g. with standard objects
• CPC Class Codes: G01N 29/4436 with a reference signal amplitude comparison G01N29/48

Abstract

A method and apparatus for the ultrasonic inspection of acoustically "noisy" specimens, such as those having intrinsically coarse grain structure (e.g., turbine rotor parts made of superalloy compositions) employs several types of ultrasonic diagnostic measurement techniques, including at least: "Pitch-Catch" analysis and "Pulse-Echo Axial Long-Wave" analysis. When a Pitch-Catch technique is utilized, a "through-transmitted" sound wave is acquired for each specimen tested and used as a calibration standard to normalize ultrasonic inspection test data for that specimen prior to any automated defect/flaw detection and sizing processing. The through transmitted sound wave is also used to equate the amplitude of "indications" (i.e., significant detected ultrasonic reflections) to that of a standardized theoretical reflector of known size (e.g., an equivalent "flat bottom hole" or EFBH reflector) so that the test sensitivity or "size of indications" can be uniquely determined for each specimen on which the test is performed. When a Pulse-Echo technique is utilized, a through-transmitted sound wave is not used (since this technique employs only a single transducer) and inspection data is instead normalized by a "back wall" reflection signal. Once ultrasonic inspection test data is obtained and digitized, a "rolling average" of the data is used to estimate the localized noise level within the material under test. For every transducer pulse data point obtained, a unique signal-to-noise (S/N) ratio is computed. Variations in S/N ratio for the material are measured at each axial slice of material volume inspected. A function of the standard deviation of the S/N values is then used to automatically adjust a detection amplitude threshold to optimize the flaw detection process and minimize false alarms.

Claims

1. In a nondestructive testing system for performing ultrasonic inspections, including a processor for controlling the positioning of at least one ultrasonic transducer element with respect to a surface of a specimen under test, acquiring, digitizing and preserving ultrasonic measurement data produced at least in part from internal scattering of ultrasonics signals in the specimen, and performing calculations based on said data, a method for dynamic automatic adjustment of a flaw/defect detection threshold, comprising the steps of:

(a) scanning a surface of a specimen under test in a predetermined pattern with at least one transducer and operating at least one transducer that is used for said scanning to acquire groups of ultrasonic measurement signal data corresponding to a plurality of ultrasonic test data points arranged within distinct planes, said distinct planes defined substantially parallel to the scanned surface of said specimen;

(b) calculating a unique signal-to-noise (S/N) ratio value corresponding to each test point of said plurality of ultrasonic test data points in a particular plane of said distinct planes based on a rolling average of signal levels measured over an area localized about each of said plurality of ultrasonic test data points acquired in step (a); and

(c) setting a flaw/defect detection threshold level for said particular plane in said specimen based on a predetermined function of a standard deviation value of all S/N values calculated for said particular plane in step (b).

2. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 1, wherein said scanning step is performed at a predetermined transducer velocity with respect to the specimen surface and using a corresponding ultrasonic pulse repetition rate that insures a plurality of ultrasonic pulse echo responses for each test point of said plurality of ultrasonic test data points scanned in said specimen.

3. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 2, wherein said ultrasonic pulse repetition rate results in at least eight pulses for each test point of said plurality of ultrasonic test data points scanned.

4. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 1, wherein during said scanning in step (a), at least one transducer is operated to provide ultrasonic pulses at a repetition rate sufficient to produce a plurality of ultrasonic pulse echo responses from a flaw/defect such that an echo-dynamic response pattern of echo response amplitudes obtained during scanning of a flaw/defect will comprise values ranging in magnitude from least to greatest by approximately a factor of two.

5. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 1, wherein during said scanning in step (a), at least one transducer is operated to scan test points in said specimen using a pulsed ultrasonic beam such that said pulsed ultrasonic beam intersects a plane of test points being scanned substantially at a 6 dB beam-waist intersection width of said ultrasonic beam.

6. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 1, wherein said calculating step includes dividing a measured signal amplitude for a single test point of said plurality of ultrasonic test data points by a local noise level measured as said rolling average of signal levels proximate said single test point.

7. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 1, wherein during said calculating step said rolling average is calculated using said measurement data acquired from test data points located within a single plane of said distinct points and acquired along at least one segment of said predetermined path scanned by at least one transducer.

8. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 7, wherein said rolling average is calculated using said measurement data acquired, proximate but not adjacent to a test point of said plurality of ultrasonic test data points, along a pair of predetermined scan path segments, a first segment of said pair preceding and separated from said test point by at least one half of a 6 dB beam-waist diameter and a second segment of said pair following and separated from said test point by at least one half of a 6 dB beam-waist diameter, such that a 6 dB beam-waist of an ultrasonic measurement signal pulse will not interfere with said rolling average calculation.

9. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 8, wherein said first and second segments are each three inches in length.

10. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 7, wherein said rolling average is calculated according to the following equation:

##EQU6## where, N_i =local noise level based on rolling average of measured signal levels proximate a test point i,

S_j =signal level at test point j,

b=6 dB beam width intersection in scan direction,

x=single scan path segment length in inches, and

n=total number of data points in scan path segment pair.

11. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 1, wherein said setting step (c) includes calculating a standard deviation value, σ

, of all said S/N ratio values in said particular plane and defining a flaw/defect detection value for setting said flaw/defect detection threshold for conducting Pitch-Catch type diagnostics as substantially equal to;

1.0+2σ

.

12. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 1, wherein said setting step (c) includes calculating a standard deviation value, σ

, of all S/N ratio values in said plane and defining a flaw/defect detection value for setting said flaw/defect detection threshold for conducting Axial Long Wave type diagnosis as substantially equal to;

1.3×

(1+2σ

) for a Near Field zone, and

1.1×

(1+2σ

) elsewhere.

13. The method according to claim 1, wherein a detected flaw/defect is indicated as valid whenever a measured signal level at a test point is equal to or greater than twice said flaw/defect detection threshold set in step (c).

14. The method according to claim 1 wherein said specimen is a turbine rotor forging and wherein said predetermined pattern in said scanning step comprises an arcuate path.

15. The method for dynamic adjustment of a flaw/defect detection threshold according to claim 1, wherein said setting step (c) includes further includes setting a flaw/defect detection threshold for each distinct plane of ultrasonic measurement data acquired in step (a).

16. An automatic flaw/defect indication and dynamic detection threshold adjustment technique for ultrasonic inspection of acoustically noisy materials, comprising the steps of:

(a) scanning a surface of a specimen under test in a predetermined pattern with at least one transducer and operating said transducer to acquire groups of ultrasonic measurement signal data corresponding to a plurality of ultrasonic test data points arranged within a plane within said specimen;

(b) calculating a unique signal-to-noise (S/N) ratio value corresponding to each test point in said plane based on a local noise level proximate to each test point wherein said local noise level is computed as a rolling average of ultrasonic measurement signal levels proximate each test data point of measurement data acquired in step (a);

(c) defining a flaw/defect detection threshold level for said plane based on a predetermined function of a standard deviation value of all S/N values calculated in step (b);

(d) repeating steps (a) through (c) for each plane inspected within said specimen, whereby said flaw/defect detection threshold level is dynamically reset for each plane;

(e) indicating a flaw/defect condition at a test point whenever a measured signal level at said test point is substantially equal to or greater than twice said flaw/defect detection threshold set in step (c).

17. The method according to claim 16 wherein said defining step (c) includes calculating a standard deviation value of all S/N ratio values in said plane and defining a flaw/defect detection value, σ

, for setting said threshold for conducting Axial Long Wave type diagnosis as substantially equal to;

1.3×

(1+2σ

) for a Near Field zone, and

1.1×

(1+2σ

) elsewhere.

18. The method according to claim 16 wherein said defining step (c) includes calculating a standard deviation value, σ

, of all S/N ratio values in said plane and defining a flaw/defect detection value for setting said threshold for conducting Pitch-Catch type diagnostics as substantially equal to;

1.0+2σ

.

19. The method according to claim 16 wherein said rolling average is calculated using measurement data acquired, proximate but not adjacent said test point, such that a sonic beam-waist of an ultrasonic measurement signal will not interfere with said rolling average calculation.

20. The method according to claim 16 wherein said specimen is a turbine rotor forging and wherein said predetermined pattern in said scanning step comprises an arcuate path.