Improving onset picking in ultrasonic testing by using a spectral entropy criterion.

Area of Science:

• Materials Science
• Acoustics
• Signal Processing

Background:

• Ultrasonic testing derives material properties from time-of-flight (ToF).
• Traditional ToF calculation methods using signal peaks are unreliable in complex materials.
• Existing onset picking methods like the Akaike information criterion (AIC) have limitations.

Purpose of the Study:

• To develop a more accurate and robust method for determining ultrasonic signal onset times.
• To address the limitations of the Akaike information criterion (AIC) picker in ultrasonic testing.

Main Methods:

• A novel spectral entropy criterion (SEC) picker was developed for ultrasonic through-transmission measurements.
• The SEC picker models ultrasonic signals within the AIC framework, considering spectral properties.
• Synthetic and experimental data were used to compare SEC and AIC picker performance.

Main Results:

• The proposed SEC picker demonstrated improved accuracy in determining signal onset times compared to the AIC picker.
• The SEC picker is less dependent on arbitrary parameters and accounts for signal spectral characteristics.
• Enhanced accuracy was particularly noted for densely sampled ultrasonic data.

Conclusions:

• The spectral entropy criterion (SEC) picker offers a more reliable approach for ultrasonic onset time determination.
• This advancement is crucial for accurate material and structural property analysis in challenging ultrasonic testing scenarios.
• The SEC picker provides a robust alternative to conventional methods, enhancing the precision of ultrasonic measurements.