Self-sensing multifunctional composite has sensing function using electrical resistance changes. Carbon Fiber Reinforced Polymer (CFRP) composite is one of the self-sensing multifunctional composites. For the reliability of the self-sensing, electrical contact between the lead wire and the carbon fibers is the most important issue. The present study focuses on the effect of the cyclic loading of lower applied strain range than the fatigue damage level. As a result, the electrical contact resistance at the copper electrode increased with the increase of cycles. That means that the electrical change at the electrodes must be considered for the long-term self-sensing monitoring system. When a four-probe method is used to measure the electrical resistance, the contact resistance effect is minimized. Moreover, angle-ply laminates have plastic deformation caused by shear loading, and that causes electrical resistance decrease during the cyclic loading. Cross-ply laminates of CFRP composites have no electrical resistance increase without damage. Quasi-isotropic laminates of CFRP composites, however, have electrical resistance decrease with the increase of the number of cycles because of the plastic deformation of the angle-ply laminates.
Carbon Fiber Reinforced Polymer (CFRP) is applied to aerospace structures and automobile structures because of its light weight, high strength and high stiffness. A delamination crack of a CFRP laminated structure is, however, quite difficult to detect for visual inspection. That causes the requirement of self-sensing CFRP structures: the self-sensing CFRP uses the reinforcement carbon fibers as sensors [1-22]. The method detects the defects from the electrical resistance change of the CFRP structure.
Schulte et al. [
Todoroki et al. [
In the present study, at first, a temperature compensation method is shown to distinguish the effect of the temperature change from the long term cyclic loading test. A new three-probe method is proposed here to measure change of electric contact resistance at an electrode: the three-probe method is an improved method made from the four-probe method. Low cyclic strain is applied here to prevent fatigue damages. The electrical resistance change is measured under the low applied strain without damages.
The material used in the experiments is unidirectional prepreg sheet (Mitsubishi Rayon Japan, PYROFIL#380). The stacking sequences are quasi-isotropic [0/45/-45/90]S and cross-ply [02/902]s. Plates of 250 mm length and 150 mm width of 1.7 mm thickness were made using an autoclave.
The pre-cure condition was 85˚C × 2 hours, and the main curing condition was 135˚C × 3 hours under 0.7 MPa pressure. After the curing, fiber volume fraction Vf was measured with the resin burn-off method, and the measured Vf was 67%. From the CFRP plates, rectangular plate specimens of 230 mm length and 12.5 mm width were produced as shown in
To measure the electrical resistance of the specimen,
the four-probe method is adopted here and a LCR meter #3522 made by Hioki Co. Japan is used. For the measurements, alternating current of 450 Hz of 30 mA is applied, and the impedance using the alternating current is measured. As the phase angle of the measured impedances of the alternating current were zero, the measured impedance was considered to be an electrical resistance here. To measure the electrical contact resistance at the electrode, a newly developed three-probe method was applied here. The three-probe method is described in the later section. The single specimen side surface was polished and the damage between the electrodes was observed here.
To obtain electrical resistance changes caused by temperature change, a hot plate and a cooler box were prepared. K-type thermo-couples were used to measure temperature of the specimen. At the middle between the electrodes, the K-type thermo-couples were mounted shown as dots in
The four-probe method is shown in
In the present study, at first, the electrical resistance of the specimen (R) is measured with the four-probe method. After that, the four-probe circuit is switched to the threeprobe circuit as shown in
For the cyclic loading test, a closed-loop hydrostaticservo material-testing machine is used. Tension-tension cyclic loading tests of stress ratio R = 0.1 are performed here. The applied maximum strain for the quasi-isotropic specimen is 2800 μ that is only 20% of the tensile fracture strain (fracture stress is 695 MPa). The frequency is 50 Hz and the sine wave is used for loading. The maximum number of cycles is 106 here.
The reason why the applied strain is set to 20% of the fracture strain is as follows. For the actual aircraft design, the most important criterion is the compression-after-impact strength, and this demands the limitation load of 20% - 30% of the compression strength. Generally, the compression strength is almost equal to the 70% of the tensile strength. This indicates that the 20% of the tensile load is approximately used as the maximum applied load for the actual composite structures. The cyclic loading strain of the cross-ply laminate is also set to this. Most of the electric current flows in the surface 0˚-plies for both types of the specimens. As the electrical resistance change is mainly controlled by the applied strain as shown in the reference [
To measure electrical resistance change, a LCR meter (type 3522) produced by Hioki Co. Japan was used. For the measurements, alternating current of 450 Hz and 30 mA was adopted, and the impedance changes were measured. As the phase angles of the measured impedances were all 0, the measured impedances were treated as being equal to electrical resistances in the present study.
For the measurement of the electrical resistance changes of the specimen, the four-probe method was used and the three-probe method was used to measure the electrical contact resistance change at electrode. Using a micro computer, switching from the four-probe to the three-probe method was performed with the interval of 40 s. This enables us to measure the electrical resistance changes of the CFRP and electrical contact resistance changes automatically. To measure the applied strain and residual strain under the completely unloaded condition, a normal type two-axis strain gage is used here. Six thermo couples were used to measure the temperature change and the reference temperature is decided at the mean applied load by averaging the six temperature values. The obtained experimental results are all transferred to the data at 20˚C using the temperature compensation. The specimen side surface was observed in situ during the cyclic loading using a video microscope.
The electrical resistance change caused by temperature change has the linear relationship. Using the linear rela-
tionship, the specimen electrical resistance can be compensated against the temperature change when the temperature distribution is measured. The electrical resistance of a CFRP specimen depends on the temperature, the damage, the applied strain and the degree of water absorption. In the present test, the specimen has no damage as well as no moisture absorption. The electrical resistance change due to applied strain in small applied load is linear relationship [
ing measures such as glass wool and using an air condition during cyclic loading tests. The temperature compensation, however, is important for the condition of large temperature change in the actual CFRP structures.
Figures 6(a) and (b) show the electrical resistance change of the quasi-isotropic and the cross-ply laminates during cyclic loading. The abscissa is the number of cycles and the ordinate is the electrical resistance change ratio using the reference resistance at 103 cycles and 20˚C. Two specimens were tested and both showed similar results. As mentioned before, the maximum strain of the cyclic loading is 20% of the fracture strain. After the cyclic loading up to 106 cycles, the polished specimen side surface was carefully observed using a video-microscope. As shown in
Figures 8(a) and (b) shows the results of the electrical contact resistance at electrodes measured with the threeprobe method. The abscissa is the number of cycles and the ordinate is the measured electrical contact resistance at electrodes. Figures 8(a) and (b) show the results of the two specimens. Two specimens were tested and the similar results were obtained.
In both laminates, rapid increase of the electrical contact resistance is observed over 105 cycles. This indicates
that the electrical contact damage caused by cyclic loading affect the electrical contact resistance over 105 cycles even if the copper plating electrodes are used. As shown in
the surface of 0˚ ply. The applied strain range in the longitudinal direction is the same for the both cases. The difference of the electrical contact resistance changes of theses laminates depends on the difference of the deformation in the transverse direction. This is discussed in the next section.
In the electrical resistance change of the cross-ply laminate in
Compared with the cross-ply laminate, the results of the quasi-isotropic laminate decrease significantly as shown in Figures 5 and 6(a). The only difference between the two laminates is the existence of the ±45˚ plies. In the present study, therefore, tensile tests and cyclic loading tests of ±45˚ laminates were performed to clarify the effect of the ±45˚ laminates in the electrical resistance change.
As mentioned before, the difference between the electrical resistance changes of the two kinds of laminates (cross-ply laminates and the quasi-isotropic laminates) is the existence of ±45˚ plies. Although the mechanical behavior of the ±45˚ laminate is different from the quasiisotropic laminate, the electrical behavior of the ±45˚ plies is identical to the ±45˚ plies in the quasi-isotropic laminate because the electric conductance of the ±45˚ plies is the same in the both laminates (quasi-isotropic laminates and ±45˚ plies laminates). In this section, therefore, ±45˚ laminate specimen is used to investigate electrical resistance change during tensile cyclic loading of ±45˚ plies.
and the transverse strain using the least square-error method. The first cycle means the first result at the completely unloaded condition. The results show that the tensile test of the ±45˚ laminate has large residual strain even in the small tensile strain in the both direction. The residual transverse strain is compression strain and the magnitude of the compression strain is larger than the residual tensile strain in the longitudinal direction.
bols are the measured results. With the increase of the residual strain at the unloading, the electrical resistance at completely unloading condition decreases significantly. Let us consider the case of the electrical resistance decrease of the ±45˚ laminate of the completely unloaded condition.
For the tensile test of the ±45˚ laminate, in-plane shear strain is applied in the plies. The electrical resistance changes during shear loading of a single ply have already reported in the reference of [
Let us estimate the effect of biaxial loading. Let the longitudinal strain of the specimen is ε1, and the in-plane transverse strain is ε2. As shown in
The strain of the fiber direction εL that inclines to the specimen longitudinal direction by 45˚ and the transverse strain εT can be obtained using the strain transformation as follows.
Substitution of the Equation (2) into the Equation (3) gives
Let us consider the case of ε1 = 400 μ. In the case, εL = −20 μ and εT = −20 μ. Using the electrical resistance change matrix caused by loading [
The Equation (5) tells (ΔR/R)L = −58 μ.
The electric current of the ±45˚ laminate flows in the fiber direction: the electric current in the 45˚-ply flows in the 45˚ direction. To visualize the electric current in the each ply, 3-D FEM analysis was performed here using commercially available FEM code ANSYS. For the FEM analysis, the electric conductance measured by Abry et al. [
electric current flows in the second ply (−45˚-ply) to the −45˚ direction. This process continues until the flow reaches to the other electrode D. This process is illustrated in
A small dent in the thickness direction in a CFRP plate brings significant decrease of the electrical resistance in the thickness direction in the reference [
Generally, CFRP laminates have angles plies such as ±45˚-plies, the results of the electric current during cyclic loading of the quasi-isotropic laminate indicates that the reference electrical resistance R0 decreases with the increase of cyclic loading even when the loading does not cause damage. This shows that the simple monitoring with the electrical resistance measurement of the CFRP structures is not applicable for the actual CFRP structures.
For the long term monitoring, we have to estimate the reference resistance R0 before monitoring as shown in the reference [
The difference of the residual strain in the transverse direction between the cross-ply and the quasi-isotropic laminates was found here. This difference caused the difference of the damage of the electrical contact resistance in
1) A temperature compensation method is shown, and the method is applied to the long term cyclic loading test to distinguish the effect of the cyclic loading from the temperature change.
2) Electrical contact resistance at electrodes made by copper plating increases with the increase of number of cycles for the quasi-isotropic laminates even when the applied loading is small enough to prevent damage. The residual strain in the transverse direction caused by the in-plane plastic shear deformation makes this damage at electrodes.
3) For the quasi-isotropic CFRP laminate, the plastic in-plane shear deformation of the angle plies causes the electrical resistance decrease during cyclic loading even for the small loading that does not cause damage. The plastic in-plane shear deformation caused residual elongation in the longitudinal direction. The residual elongation means the decrease of the thickness, and this caused the electrical resistance decrease for the quasi-isotropic CFRP laminate.