Damage detection of carbon reinforced composites using nondestructive evaluation with ultrasound and electromagnetic methods

CFRP have applications among most different domains due their low density, high elastic modulus and high ultimate strength along the carbon fibers direction, no fatigue and the expansion coefficient is small. This paper presents the behavior of carbon fiber woven-PPS composites at low velocity impacts. The transversal electrical conductivity is modified due to the plastic deformation following the impacts, and thus electromagnetic procedures can be used for assessment of CFRP using a high resolution sensor with metamaterials lens and comparing the results with those obtained from ultrasound testing with phased array sensor. The area of the delamination is overestimated when the method of phased array ultrasound is used and substantially underestimated by the electromagnetic testing. There were a good agreement between the simulations with finite element method and experimental measurements.


Introduction
Carbon fiber reinforced plastics composites are lightweight, strong materials used in the manufacturing of numerous products, from the sports goods to aerospace industry components. CFRP allow the production of goods without scrap parts, becoming a cost effective technology [1]. In general, CFRP composites use thermosetting resins such as epoxy, polyester, or vinyl ester. High performance thermoplastic resins and more particularly Polyphenylene sulfide (PPS) [2] represent a promising alternative to thermosetting resins problems because it offers a number of advantages over epoxy resins (chemical and impact resistance over a wide range of temperatures, high hardness and stiffness and superb long-term creep under load properties, extremely low moisture absorption).
For 2020, it is estimated that the carbon fiber demand will outstrip supply, doubling the production of 2015 (tripling the production of 2010). In order to avoid the incurring of all costs within selection between good and bad parts, NDE is required from the manufacturing stage of composites, from base fiber and production of matrix. Nondestructive evaluation is no longer about "pass/fail" [3]. NDE should provide validation of ply sequences and fiber/woven orientation, the state of the matrix, parameters of the final materials. Low velocity impacts (LVI), that can be defined by quasi-static events of which superior limits can be in the range of tens of m/s [4], depending on material properties as stiffness, can produce delaminations, carbon fiber breaks, cracking of matrix [5] etc.
This paper present the possibility to use a high resolution sensor based on metamaterials lens to assess the delaminations in CFRP with Polyphenylene sulphide matrix reinforced with carbon fibers woven fabric, created by LVI and comparing the results with those obtained from top scan ultrasound using phased array method.

Studied samples
The plates taken into study have the dimensions of 150mm×100mm×4.2mm [17]. The composite, named CETEX, is made from PPS matrix reinforced with 12 layers of carbon fibers 5Harness satin fabric type. The carbon fibers are T300JB with 1.75 g/cm 3 density, fabric surface density is 285g/cm 2 and overall fiber volume fraction 50 ± 3% [18]. In figure 1   The samples were impacted with energies of 2 J, 4 J, 6 J, 8 J, 10 J, and 12 J at room temperature using equipment FRACTOVIS PLUS 9350-CEAST-Instron USA with a hemispherical bumper head having 20 mm diameter and 2.045 kg weight, according to ASTM [17]. The intermediate energy of 10J has been taken into consideration before the failure of the fibers that have taken place at 12J.

Mechanical tests
The conditions of impacts were established by simulating the Drop Weight impact method [19] using Finite Element Method (FEM). Drop weight testing is the preferred method when using low velocity impact testing and the results may also be enhanced with post impact testing. The energy of impact  (1) with δ the displacement of the plate in point i; F(t) is recorded by the data acquisition system, M total is total mass of the impactor. The absorbed energy until the i point is calculated as the area described by The speed till point i is calculated by simple integration of Force-Time diagram

Electromagnetic testing
The plates were examined using the sensor with metamaterial lens for detection. The sensor has the emission part made from a 20mm diameter coil, 2 turns of Cu wire (1.2mm diameter). The reception coil has a 1mm diameter, 2 turns of Cu wire (0.1mm diameter). In front of the reception coil is placed a conductive screen made from a flexible foil LONGLITE TM 200 produced by ROGERS CORPORATION USA, having a circular aperture, with 0.1mm diameter, coaxially placed at 0.05mm distance from the reception coil. The screen was grounded. The circular aperture serves for the diffraction of the evanescent waves, appeared in the slits. A conical Swiss roll [20] is placed above the reception coil and acts as metamaterials allowing the transmission of the evanescent waves [20] from the region between carbon fibers towards the reception coil ( Figure 4). The transducer was fixed, being connected to the Network/Spectrum/Impedance Analyzer 4395A Agilent USA. The plate is displaced by a motorized XY stage Newmark USA, commanded by PC through RS232 interface. The command and the data acquisition were made through a GPIB IEEE 488 interface, using programs elaborated in Matlab 2014a. The surface was raster scanned using 1mm steps in both directions.

Ultrasound testing
The testing using ultrasound phased array sensors has been effectuated using Phasor XS -GE USA. The sensor contains a linear array of 32 sensors having the pitch of 0.5mm and the central frequency 5MHz. The sensor is placed on an edge of 36° and the displacement has been assured by an axial scanned ENCSTD with linear encoder ( Figure 5). The coupling is a gel with average viscosity.  Figure 5. Ultrasound testing.

Experimental results and discussions
The composite specimen prepared for mechanical tests, is cropped from the central region of a sample impacted with 10J. This impact only produce a dent. The tensile tests were done using Instron E8801 with a displacement speed of 0.2mm/min, using the pressure of 70barr [21]. In figure 6 is presented the residual stress at tensile test after impacts. Approximating the plot as a parabola, it can be shown that the residual stress of the sample impacted with 12 J energy is 50% from the non-impacted samples.  Taking into account the values of displacement at impact simulated by FEM in the conditions specified above, there were found a good concordance with the values experimentally determined during the impact tests ( Figure 7). The tests were determined on the sample impacted with 10J, energy below the energy where breaking of the fiber appears [22].
Microscopic analysis using Olympus Microscope BX51 of the composite specimens impacted with 10J has been effectuated for evaluation of the the damage initiation and propagation of delamination. The specimen has been cut on the middle of the impacted zone and prepared for optical inspection. The damage developed in the 5-HS fabric under impact takes the form of weft yarn cracking, which is perpendicular to the loading direction. The damage is developed at the yarn crossover positions, and then propagates into the matrix (Figure 8). b. Figure 9. Scanning of the plate impacted with 10J energy: a) ultrasound; b) electromagnetic image.
The obtained experimental results show that the area of delamination is overestimated by the proposed ultrasound method and it is underestimated by the electromagnetic. The impact damage exists at different levels through thickness. The delamination near the back wall is larger than the ones near the front wall (where the impact took place), there is rapid local bending at the impact point and at the face of the delamination, almost all the energy will be reflected back to the sensor as an echo. It is not possible to penetrate past the first incident delamination to those in its shadow; all the energy is then confined between the first incident delamination and the front wall to create repeat echoes. Multiple level delaminations are detected by virtue of the conical damage feature (those near the back wall being larger in area).The underestimation using electromagnetic method is given by modification of transversal electrical conductivity of the composite material due to the impact, that took place in the central region of the impact, where the material suffers a plastic deformation. The electromagnetic sensor "senses" only the subsurface delamination.

Conclusions
CFRP plates made from 12 layers of carbon fibers 5Harness satin fabric embedded into PPS matrix have been impacted with different energies in order to study their damage using nondestructive evaluation methods by means of ultrasound phased array and electromagnetic sensors. The residual stress of the sample impacted with 12 J energy is 50% from the non-impacted samples. The area of delamination is overestimated by phased array ultrasound method because the delamination on the back face is larger than the ones on the impacted face and due to repeated back end echoes and respectively is underestimated by the electromagnetic method due to plastic deformation that modify local conductivity. In order to obtain more eloquent results, a fusion of data obtained by both nondestructive evaluation methods using Dempster Shafer theory of evidence is required and will be further studied.