Carbon fiber (CF) laminates are a class of advanced composites with distinctly anisotropic properties at multiple scales, from the single CF strand to the fully engineered 3D part. There are many applications of these laminates in the automobile, aerospace, and energy storage industries primarily due to their high strength-to-weight ratio. These laminates are comprised of several ply layers arranged in different angular orientations (depending on the thickness and mechanical requirements of the completed part) all embedded in a resin matrix. During the curing process, a strong adhesion between the fibers and ply layers occurs which provides strength and structural integrity to the part. Due to the multitude of steps involved to create a layered CF composite with specific mechanical properties and shape, there are many opportunities for defects to occur[1]. Defects on the meso- to macro- scales (often the last sign before structural failures), such as cracks, fiber or bundle breakage, and delamination are of significant interest as monitoring targets for information on a composite’s health. These defects often occur deep within the internal CF structure and are not visible from the outside surface, thus necessitating some method for analyzing the interior of a carbon fiber laminate post cure.
Current internal defect detection methods include ultrasonic testing (UT), laser shearography, modal acoustic emission (MAE), eddy current testing (ECT), and the “gold standard” Xray computed tomography (Xray CT). Each of these techniques has its own unique advantages and limitations in terms of the ability to provide structural information or detect defects within CF laminates.
UT can be split into two main categories: air-coupled UT and gel/water-coupled. Both rely on the efficient coupling and propagation of ultrasonic waves into the material under test, where the reflection or transmission (depending on the mode) of these mechanical waves is monitored. Air-coupled UT is limited to accurate inspection of CF laminates to only a few millimeters below the surface and is difficult to implement on structures with rough surfaces, such is the case for many of these structures. The detection depth can be increased using water or gel couplants, but this quickly becomes impractical for large structures, since they either need to be completely immersed or have a constant supply of couplant (e.g. water or gel), which is costly and messy. Furthermore, UT has increasing difficulties with multi-ply laminates due to impedance mismatches at the many interfaces, wave scattering effects due to fibers, and absorption by the resin matrix which requires the use of lower frequencies[2]. These lower frequencies again limit the penetration depth to a few millimeters, even with the use of couplants[3]. Lastly, “shadow effects” are known to severely limit the sensitivity to defects that are found underneath shallower surface defects.
Laser shearography is a large area imaging technique that can provide accurate identification of the size and location of defects based on differences in localized material strain rates measured optically from the CF composite surface. Nonetheless, this technique is typically limited to inspecting parts that are relatively thin (often < 10 mm) and cannot provide any information about defect location in terms of where in the depth (i.e. through-the-thickness) the defect is located.
MAE is an expansion of classic acoustic emission (AE). Both are mechanical guided wave techniques that identify defects by their characteristic acoustic signals. These acoustic signals are actively induced through applying stress to the structure under test and then “listening” to the emitted waveforms via attached (in-contact) transducers. MAE utilizes the differences in “modes” within the acoustic emission waveform to deconvolute the contributions of different events which can be used to understand the damage type and location within composite structures[4]. MAE does not provide imaging capabilities and is used more often as a pass/fail technique for QA/QC. MAE is an inherently time-consuming process due to the need to actively attach transducers to the CF laminate in specific locations and the need for calibration on the part being tested to obtain accurate detection capabilities. For large CF composite parts, MAE can require several hours or more to set up and perform the inspection. For Type III or Type IV carbon overwrapped pressure vessels (COPVs), analysis times are extended significantly longer due to the need to pressurize the tank during the MAE inspection, while also presenting a safety hazard. Each AE “event” is discrete and unique in that they record individual stress events that cannot be stopped, only occur for a finite period of time, and cannot ever be identically reproduced, which makes characterization of unknowns difficult[3].
Xray CT provides a full 3D image of the CF laminate that carries high resolution depth-resolved information about the structure and can be used to identify almost any defect type that provides a change in the density. Nonetheless, Xray CT requires costly instrumentation and highly trained operators, while also being limited in the overall size of object that can be analyzed. Moreover, Xray CT is currently a laboratory technique and has yet to be made field portable. For large specimens, such as the CF overwrap on an 86 L Type III or 90 L Type IV COPVs, analysis times are quite long and can vary from 8–16 hours.
Eddy current testing (ECT) is a non-contact characterization method that relies on the generation of localized eddy currents in a conducting material using an AC electromagnetic field (EMF). The primary transmit field (B1) induces eddy currents, which generate a secondary magnetic field (B2) that interferes with the B1 field. The result of this interference is a change in impedance of the transmit coil. Conventional ECT methods suffer from shallow penetration depth at high frequencies and low sensitivity to structural changes at low frequencies, which has made accurate measurement deeper than ~ 4 mm unachievable in CF laminates[5–12].
Due to the heterogeneous and inherently complex nature of CF composites, fast, reliable cost-effective defect detection has not yet been realized throughout parts greater than a few millimeters thick. This leaves a need for non-destructive evaluation (NDE) techniques to be developed for identification and characterization of deep-lying defect within CF laminate structures. In this paper we describe a new carbon fiber NDE, which we’ve termed Electromagnetic Inductive Coupling Analysis, or EMICA.
1.1. Electromagnetic-Inductive-Coupling-Analysis (EMICA)
The EMICA approach has been recently developed[13]. The measurement approach includes a sensor coil (solenoid or pancake form factor) broadcasting a discrete low-frequency electromagnetic field (EMF) into the CF specimen under test. Due to the conductive nature of the carbon fiber, an inductive coupling is established between the coil and carbon fiber. The electrical components of the sensing circuit are incredibly stable, resulting in a high sensitivity to changes in the CF laminate itself due to changes in the physical structure of the CF laminate (i.e. defects like cuts, delaminations, and impact damage).
The excitation waveform is output from the sensor coil and returns with information on structural features within the CF laminates (i.e. changes in the reflected power of the sensor coil). As we will show, EMICA takes advantage of the flow primary B1 field flow into the carbon fiber, achieving something similar to an electrical connection into each fiber, but in a non-contact way. EMICA probes both the material’s surface and depth simultaneously. The highest eddy current intensities are often found at internal orthogonal ply layer interfaces (not at the surface), and the rate of decay of the EMF through-the-thickness is significantly lower, i.e. EMICA frequencies reach significantly further into a conductive material than would be expected using standard penetration depth calculations. EMICA is the electromagnetic equivalent of performing an ultrasonic shear and compression wave experiment simultaneously, point-by-point across a specimen.
In this article, we establish an empirical understanding of EM wave propagation through “healthy” complex 3D CF laminate structures using EMF transmit/receive maps as a function of excitation frequency. These studies inform selection of the operating frequencies used for EMICA. Because of the complexities of modeling electromagnetic field in carbon fiber laminates, we designed and fabricated a library of “standard reference materials” (SRM) library containing 6 mm thick flat CF panels with healthy and known defects. This library was used to empirically evaluate the EMICA imaging capability as a function of frequency and experimentally optimize the data collection protocol for detection at depth. We then extended what we learned in 6 mm thick panels to EMICA imaging of a ½” (12.7 mm) flat panel having three known defects buried inside the panel at depths of 3 mm, 6 mm, and 9 mm.