Through transmission ultrasonic inspection of fiber waviness for thickness-tapered composites using ultrasound non-reciprocity: Simulation and experiment

Ultrasonic detection of wrinkles in composites with gradual phase shift migration

Fiber reinforced polymer composites (FRPs) are essential for various industrial fields, but wrinkles inside will greatly reduce their mechanical properties. Full-matrix capture (FMC) is a popular data structure for ultrasonic phased array imaging in composites. However, such structure may lead to data redundancy and noise interference. In this paper, a gradual phase shift migration (GPSM) is proposed to characterize wrinkles accurately. The gradual matrix is formed from the expansion along principal diagonal of FMC data with equal transmitter-receiver spacing. The dilemma between lateral resolution and sidelobe interference intensity is resolved to obtain the best imaging resolution by selecting an appropriate data structure. Moreover, to address the inconsistency of ultrasound velocities at different propagation directions caused by anisotropy of composites, the angle-dependant velocity is corrected by backwall reflection method (BRM). Based on gradual matrix data, the velocity-corrected phase shift factor is applied in the GPSM algorithm to obtain the wavefield at different depths through a layer-by-layer wavefield extrapolation. The experimental results indicate that four wrinkles can be detected in thick hybrid carbon-glass FRPs based on GPSM, with angle detection errors less than 6%. Furthermore, the GPSM method combining partial diagonal data takes only 0.5 s, achieving 60% improvement in computational efficiency compared to that with all gradual matrix data. The proposed method can be applied for high-resolution imaging of various multilayered medium in real-time.

Advanced Ultrasonic Inspection of Thick-Section Composite Structures for In-Field Asset Maintenance

An investigation into the inspection capabilities of in-field advanced ultrasound detection for use on ultra-thick (20 to 100 mm) glass fibre-reinforced polyester composites is presented. Plates were manufactured using custom moulding techniques, such that delamination flaws were created at calibrated depths. The full matrix capture technique with an on-board total focussing method was used to detect flaws scanned by a 0.5 MHz linear array probe. Flaw through-thickness dimensions were altered to assess the threshold for crack face separation at which delaminations could be identified. Furthermore, part thickness and in-plane flaw dimensions were varied to identify the inspection capability limitations of advanced ultrasonics for thick composites. The results presented in this study demonstrate an inverse relationship between the ability to find delaminations and plate thicknesses, with inspections successful at depths up to 74 mm. When the delamination thickness exhibited surface-to-surface contact, the inspection capability was reduced to 35 mm. There was an exponential decay relationship between the accuracy of the flaw depth measurement and plate thickness, likely due to the necessity of low probe frequencies. The effective inspection depth was determined to be in the range of 1 to 20 times the wavelength. It is speculated that the accuracy of measurements could be improved using probes with novel coupling solutions, and detectors with optimised signal processing/filtration algorithms.

Imaging of Fiber Waviness in Thick Composites with Unknown Material Properties Using Probability-Based Ultrasound Non-Reciprocity

Fiber waviness and voids may be produced in thick composites due to improper manufacturing conditions and consequently pose a risk of structural failure. A proof-of-concept solution for imaging fiber waviness in thick porous composites was proposed from both numerical and experimental studies, via calculating ultrasound non-reciprocity along different wave paths in a sensing network constructed by two phased array probes. Time-frequency analyses were conducted to reveal the cause of ultrasound non-reciprocity in wavy composites. Subsequently, the number of elements in the probes and excitation voltages was determined for fiber waviness imaging using the ultrasound non-reciprocity with a probability-based diagnostic algorithm. The fiber angle gradient was observed to cause ultrasound non-reciprocity and fiber waviness in the thick wavy composites were successfully imaged regardless of presence of voids. This study proposes a new feature for the ultrasonic imaging of fiber waviness and is expected to contribute to processing improvement in thick composites without prior knowledge of material anisotropy.