Single transmission phase- and frequency-modulated coded excitation for enhanced inspection of thick complex industrial components using a scalable, flexible, lead-free, ultrasonic array

Abstract

To address the growing challenge of applying ultrasonic non-destructive evaluation to complex industrial components, flexible ultrasonic arrays have emerged as a conformable solution to inspect such geometries, thereby removing the need for custom-designed wedges to conform to surfaces. Flexible lead-based arrays have been used in prior research. They offer high piezoelectric coefficient, however, they pose human health and environmental risks, and fail to comply with global initiatives including the Restriction of Hazardous Substances (RoHS) regulation enacted by the European Union. Although numerous studies on the piezoelectric properties of lead-free materials have been conducted, the uptake of technology and implementation in practice has been slow. In this work, a scalable, RoHS-compliant, flexible ultrasonic array was employed to improve operability in thick convex and concave components. However, the lead-free array exhibits lower piezoelectric coefficient compared to its lead-based counterparts, resulting in reduced signal quality. To tackle this shortcoming, single transmission phase-modulated Barker and frequency-modulated chirp excitation schemes, in conjunction with pulse compression, were employed to improve the signal quality. Subsequently, their impact was studied in terms of imaging quality, through Full Matrix Capture (FMC) acquisition methodology and Total Focusing Method (TFM) imaging, and Signal-to-Noise Ratio (SNR) measurements. A novel SNR method was presented. Existing SNR approaches evaluate image quality by calculating it within a designated area surrounding the target, where the noise level is quantified as the root mean square of the image noise, omitting any indication of the target. In addition to the noise level, artifacts from matched filters and sidelobes require quantitative evaluation. The new SNR technique was proposed to automate the selection of regions when characterising the SNR. The SNR was calculated across regions of varying size, with the region size where the SNR values converged being selected. This technique was utilised in a comparative analysis including a single-cycle pulse excitation, modulated Barker and chirp excitation schemes with equivalent energy levels in simulation and experimentally. The simulated and experimental results showed good agreement, with some discrepancies attributed to imperfections in the experimental conditions. SNR improvement exceeding 2.6 dB was observed experimentally, with the coded excitation techniques showing higher SNR and better image quality without sacrificing acquisition speed. Moreover, sidelobe artifacts were evident in all TFM images, while the coded excitation images further exhibited matched filter processing artifacts. The flexibility of the array was assessed in the subsequent two experiments to determine its effectiveness in improving operability in complex-geometry samples. The convex and concave samples pre-aligned the array to promote a converging and diverging ultrasonic beam, respectively. In all cases, the array demonstrated excellent conformity with the components, and the coded excitation schemes consistently achieved better imaging quality relative to the pulse excitation case.