Investigation of Effect of Non-Uniformities in Unidirectional Carbon Fiber Composite on the Lcr Waves Speed Using Phased Arrays

Abstract

Structural parts benefit on a reliable, nondestructive inspection technique to measure stresses, both applied and residual. Among the candidates, ultrasonic techniques have proven to have enough sensitivity to strain to be employed in service. The way to obtain the stresses is through the measurement of the time-of-flight inside the material and relates it to the strain by acoustoelastic theory or previous measurements. However, stress measurement using ultrasound strongly depends on the uniformity of the material under inspection. In composite materials, the time-of-flight is influenced by microdefects and misalignments in the fibers as well as by the applied strain and temperature. This last factor can be known and controlled, but non uniformities are a characteristic of one particular region or part. Thus, unless employed to a very particular case of a completely uniform region been inspected in a special developed part, UT could not be used to measure stresses in this kind of material without some previous information about it. This work presents an investigation about the effect of non-uniformities in carbon fiber-epoxy pre-preg composites and how to relate them with the time-of-flight of critically refracted longitudinal waves (Lcr) propagating in the fiber direction (main structural direction). A Phased Array System (PAS) with probe of 5 MHz and 64 transducers are employed to generate an image of each part in the region where the Lcr wave travels. The image is created employing the Total Focusing Method (TFM). Two bars of carbon fiber composites with epoxy matrix (HexTow® AS4 / Hexply® 8552) were tested. Five measurement positions are selected, uniformly distributed on the part surface. Statistically significant differences between the parts were found in the time-of-flight for Lcr waves when no stress is applied; even knowing they were manufactured using the same process and materials. The parts were evaluated using the PAS. No difference was found between measurements in the same bar. The parameter chosen to evaluate the non-uniformity was the peak value of the back-wall signal divided by the RMS value of the noise intensity, which was called signal-to-noise ratio (SNR). The results show also significant difference between the SNR of both parts, although with higher dispersion than with Lcr. It can be noticed that there is a correlation between the time-of-flight of Lcr waves and the SNR, indicating that the research could be extended to the development of a new joint technique to be used to measure stresses in composite parts.