With the implementation of major aircraft structures fabricated from carbon fiber reinforced plastic materials, lightning protection has become a more complicated issue for designers and engineers to solve.

Cracking of coatings and surface layers is evident on a variety of structures including buildings, automobiles, and aircraft. In some situations, the appearance of the coated or painted surface is degraded and the aesthetic appeal is lessened. However, in others, such as composite aircraft structures, paint cracking is both aesthetically undesirable and potentially deleterious from the electromagnetic effects aspect.

Figure 1. Representative surface protection scheme modeled using COMSOL Multiphysics. The composite was modeled as two layers indicated by the red and light blue regions.

In the latter case, cracks can propagate into the structure or around fasteners, providing a path for moisture and other environmental species to enter, resulting in corrosion and degradation of the protection measures including expanded metal foils (EMF) required for lightning abatement and safe operation. Consequently, over several decades there have been numerous efforts and investigations concerning the degradation of surface layer protection schemes.

Figure 2. Examples of SWD/LWD ratios from 0.25 to 1.00 for a 1 in2 EMF.
Cracking typically develops over extended periods of time due to environmental factors and thermal cycling of the surface layers and substructure. The thermal cycling of aircraft is a direct result of the typical ground-to-air-toground, often repeated, flight cycle. Subsequently, stresses accumulate in the coatings, eventually leading to failure of their protective functionality.

There are several contributors to the stress buildup, including the paint, primer, corrosion isolation layer, surfacer, EMF, and the underlying composite substructure. Boeing recently did a study that focused primarily on the EMF contribution to the cracking mechanism.

A representative surface layer protection scheme was addressed that was composed of the layers mentioned above. The approach taken was to simulate the temperature cycle of the layers using a coefficient of thermal expansion (CTE) model developed with the commercially available COMSOL Multiphysics software.

The simulation allowed determination of the thermal stress and displacements that result from repeated duty cycles. Though the full complexity of crack genesis was not included, some insight could be gained regarding what the sensitive parameters of the EMF may be and the variations that can be employed to mitigate the resulting stress and displacements that lead to cracking. Of particular interest are the EMF width, height, aspect ratio, composition (aluminum (Al) or copper (Cu)), and surface layup structure.

In the case of Al used for EMF, there is a need for fiberglass between the aluminum and the structure to prevent galvanic corrosion. Though not the major thrust of this research, the potential effect on stress and displacement that results from the glass transition temperature of the paint layer must be considered.

Model Solving

Boeing developed a CTE model to simulate the effects of EMFs incorporated in a composite surface protection scheme undergoing thermal excursions using COMSOL Multiphysics software, a finite element solver that contains a variety of physics and engineering applications with an emphasis on coupled or multiphysics analysis.

In particular, Boeing used the Thermal Stress Multiphysics Interface that combines solid mechanics with heat transfer. Coupling occurs where the temperature from heat transfer acts as a thermal load for the solid mechanics, causing thermal expansion. The interface has the equations and features for stress analysis and general linear solid mechanics, solving for the displacements.

A representative surface protection scheme was created using the COMSOL model builder. The layup consisted of paint, primer, fiberglass, surfacer, EMF (Cu or Al), and fiberglass on a composite substrate. The height and size of all the layers could be varied using input parameters. In addition to geometrical parameters, the initial and final temperatures could be varied and were chosen to be typical of altitude and ground, respectively. The strain reference temperature was the initial temperature and a representative heat transfer coefficient of 5 W/m2K was used.

Representative Input Material Parameters

In addition to the height and width, the aspect ratio of the EMF mesh could also be varied. The metallic mesh aspect ratio is given by SWD/LWD where SWD is the short way of the diamond and LWD is the long way of the diamond, as described by Dexmet, a commercial producer of EMF.

Figure 3. Von-Mises stress patterns due to thermal air-to-ground heating cycle. EMF mesh bleed-through is evident in the central portion of the figure, and relatively large displacement variations are observed above the metal and voids. Note that the displacements have been magnified for illustrative purposes.
The material properties needed for each of the layers were the CTE, heat capacity, density, thermal conductivity, Young's Modulus, and Poisson's ratio. Many of these parameters have a dependency on temperature, but for the simple model references here, researchers used the values across the temperature range of interest. The one exception was the CTE of the paint layer where a step function was employed at the glass transition temperature.

The paint had a larger CTE, heat capacity, and Poisson's Ratio than the underlying layers. Among other effects, this meant that the paint would undergo compressive stress when the layers were heated, and tensile strain when cooled. Hot materials under stress relax through creep, but this effect was not included in this particular model. For the other material parameters used, the density, thermal conductivity, and Young's Modulus were larger than the paint layer. In particular, for the EMF, Al has a larger CTE than Cu, and a smaller Young's Modulus.

Simulation Results

For the purposes of this research, Boeing confined its simulations to heating over a representative air-to-ground temperature range that the surface protection scheme was expected to perform. The resulting false color Von-Mises stress and displacement patterns are shown in Figure 3. In this figure, the view is from the top paint layer with cross-sectional views from the four sides. Obviously, the displacements have been magnified to highlight the movement that is induced by the temperature cycle. In the central portion of the figure, the pattern of the underlying EMF can be seen.

Variations in the displacement above the metal mesh and voids are quite evident in the cross-sectional profiles. Also, high stress (red—high, blue—low) can be seen in the mesh itself and the region in the mesh voids where surfacer material was modeled. The profiles show that the stress decreases from the bottom to the top of the surface layers. High stresses are clearly indicated in the EMF, where a semi-transparent stress image was generated.

A more quantitative examination of the EMF stresses and displacements could be determined by creating profiles along a selected path through the metallic layer. For this profile, the EMF was composed of Al with a SWD/LWD ratio of 0.50. It was expected that the profile would show five transitions as each metal-void region is crossed.

The arrows in Figure 4 indicate central locations where stress and displacements were determined for parametric variations of EMF SWD/LWD ratio, width, and height. These determinations were made for both Al and Cu EMFs. The nominal SWD/LWD ratio was 0.50. Fiberglass was modeled above and below both Al and Cu, but in practice is used only below the Al EMF.

For the variation of SWD/LWD from 0.25 to 0.75 for both Al and Cu, the displacement decreased slightly with an increasing SWD/LWD ratio. A higher SWD/LWD ratio corresponded to a more open mesh structure (as shown in Figure 2), resulting in lower metal density and hence lower weight. Also, inclusion of fiberglass below the Cu increased the displacement.

By varying the EMF width by a factor of 2.6 for both Al and Cu, the displacement remained essentially constant, but was significantly greater for Al than Cu. Varying the EMF height by a factor of 2.7 for both Al and Cu resulted in a displacement that increased with metal height and was also significantly greater for Al than Cu.

Boeing used a representative temperature dependence of the CTE for the paint layer to simulate the effect of a shift of the glass transition temperature, tg, from within the nominal temperature range to above it at 350 K. This variation permitted the examination of what occurs if the paint CTE remains constant throughout the nominal operating range.

There was a reduction in the surface displacement of the paint when the tg was above the maximum expected operating temperature. However, when the tg was above the operating temperature range, the paint remained in a more brittle, glassy state, which is expected to be prone to crack formation. Moving the tg below the operating range reduced the modulus that may compensate for the increased CTE that would occur. Such trade studies will be the subject of future simulations using this model.

Test Results

Figure 4. Relative stress and displacement along EMF profile path indicated in Figure 5 for Al with an SWD/LWD of 0.50. The arrows indicate the locations selected for the parametric variations to be shown below. Note that five transition regions appear as delineated by the EMF.
Quantitative determinations of stress and displacement were not conducted in the experimental evaluations. Therefore, no detailed comparisons were possible with this model. However, qualitative agreement was observed with the simulations since the EAF (expanded aluminum foil) consistently exhibited greater displacement over the various parameter sets than the ECF (expanded copper foil) displacements.

Figure 5. EMF displacement dependence on metal mesh SWD/LWD ratio for Al and Cu.
Researchers associated greater thermally induced displacements with increased probability that cracks will eventually become evident. The displacement differences may be small, but over thousands of cycles will eventually generate residual stress, defects, and result in cracking. From this standpoint, both the simulations and testing indicate that Cu would be a better choice for the EMF than Al.

The individual parametric variations also suggested some interesting effects. The parametric variation of SWD/LWD shown in Figure 5 indicated that largerratio, more open EMF meshes lead to lower displacements. The dependence is weak, but high thermal cycling has a cumulative effect. From a weight perspective, higher SWD/LWD is also desirable. Provided the EME function is not seriously degraded, there appears to be benefit with the more open mesh from multiple perspectives.

The effect of the additional layer of fiberglass under the EMF is also shown in Figure 5. When the fiberglass was added under ECF, the displacement was significantly increased. The remaining difference between EAF and ECF is most likely due to the larger CTE of aluminum by ~35%. As noted previously, the fiberglass is incorporated under aluminum to inhibit galvanic corrosion.

Examination of the thermally induced displacements suggests that there is little cracking penalty from increasing the width of the EMF. Hence, if greater current-carrying capability is desired from the electromagnetic environment protection function, increased width appears to be a viable approach. Of course, increased width leads to greater weight penalty and these conflicting requirements need to be balanced.

Alternatively, the increased current carrying capacity of the EMF layer may also be realized with increasing height. However, height increase is not as desirable as it leads to greater displacements and hence cracking likelihood.

This article is based on SAE technical paper 2013-01-2132 by Jeffrey Morgan of Boeing (Chicago, IL).

Aerospace & Defense Technology Magazine

This article first appeared in the February, 2014 issue of Aerospace & Defense Technology Magazine.

Read more articles from this issue here.

Read more articles from the archives here.