[Paper Review] Air-coupled thickness measurements of stainless steel

Grunde Waag, Lars Hoff+1 more|arXiv (Cornell University)|Oct 1, 2012

Optical measurement and interference techniques4 references8 citations

TL;DR

This study demonstrates air-coupled through-transmission ultrasound for non-contact thickness measurement of stainless steel plates using broadband pulses in the 200–600 kHz range. It resolves thickness differences of 0.2 mm (2%) between regions of 10.0 mm, 9.8 mm, and 9.6 mm, with second harmonic resonance matching expected values and first harmonic deviations attributed to finite transducer aperture effects.

ABSTRACT

A method of measuring the thickness of steel plates using through transmission of an acoustic pulse is demonstrated. This study has been done on a stainless steel plate with regions of thickness 10:0 mm, 9:8 mm, and 9:6 mm, using broadband pulses with energy in 200 kHz to 600 kHz band. Ultimately the goal is to perform similar air-coupled thickness measurements in a single sided pitch-catch measurement setup. The spectra of the transmitted pulses show the first and second harmonics of the compressional waves in the plate. When compared to a plane wave model of a fluid layer embedded in air, the second harmonic of the plate resonance fits well with the expected value. However, the first harmonic deviates such that the plate appears thicker at this resonance. This is believed to be caused by the finite aperture of the transmitting transducer, causing deviations from a plane wave. Thickness differences of 0:2 mm between the different regions of the plate were shown to be resolved. A third peak was found in the spectra. The origin of this peaks has not been verified, but is believed to come from the third harmonic of the shear wave in the steel plate.

Motivation & Objective

• To develop a non-contact, air-coupled ultrasound method for measuring steel plate thickness without liquid couplants.
• To overcome the high transmission loss (~-45 dB) between air and steel by using broadband pulses and signal averaging.
• To resolve thickness differences as small as 0.2 mm in a single-sided pitch-catch configuration using resonance frequency analysis.
• To investigate discrepancies between measured and expected resonance frequencies, particularly for the first harmonic.
• To explore the origin of an unverified third peak in the transmission spectrum, possibly linked to shear wave harmonics.

Proposed method

• Employed through-transmission ultrasound with air as the coupling medium between two transducers mounted 100 mm from a stainless steel plate.
• Used custom-built 540 kHz transducer (18 mm diameter) for transmission and NCT500-D6 500 kHz receiver (6 mm active element), both with plane elements.
• Applied linear chirp pulses from 200–600 kHz, with 200 recordings averaged to improve signal-to-noise ratio.
• Performed time-gating on the received signal (130 ms window) to isolate the transmitted pulse and reduce interference from electromagnetic coupling and reverberations.
• Estimated power spectra from time-gated signals and applied the fluid-layer transmission model (Eq. 2) to relate resonance frequencies to plate thickness.
• Used the second harmonic peak in region A (10.0 mm) to estimate the speed of sound in steel as 5720 m/s for subsequent thickness calculations.

Experimental results

Research questions

• RQ1 Can air-coupled ultrasound resolve thickness differences of 0.2 mm in stainless steel plates using through-transmission?
• RQ2 Why does the first harmonic resonance frequency yield a thickness estimate ~7% higher than nominal, deviating from the plane wave model?
• RQ3 What causes the third observed peak near 460 kHz, and could it be related to shear wave harmonics?
• RQ4 How do finite transducer aperture effects influence resonance frequency measurements in air-coupled ultrasonic systems?
• RQ5 Can inclusion of shear wave modes and non-plane wave effects improve the accuracy of thickness estimation?

Key findings

• Thickness differences of 0.2 mm (2%) between regions of 10.0 mm, 9.8 mm, and 9.6 mm were clearly resolved in the transmission spectrum.
• The second harmonic resonance (near 570–590 kHz) yielded thickness estimates within ~1% of nominal values, confirming good agreement with the fluid-layer model.
• The first harmonic resonance (near 270 kHz) produced thickness estimates ~7% higher than nominal, attributed to non-plane wave effects from the finite transducer aperture.
• The third peak at ~460 kHz could not be explained by the compressional wave model but may correspond to the third harmonic of shear waves, with estimated thicknesses of 10.16 mm, 9.93 mm, and 9.84 mm for regions A, B, and C.
• The speed of sound in the steel plate was estimated as 5720 m/s using the second harmonic in region A, enabling consistent thickness calculations across all regions.
• The method demonstrated feasibility for single-sided, non-contact thickness measurement using air-coupled ultrasound, with potential for improved accuracy through inclusion of shear wave and aperture correction models.