Picking the pattern for a stealth antenna

Altran’s team works mainly in the aerospace and defense industry, and has developed projects related to studies of antenna placement as well as radar cross-section prediction and control. One project addresses one of the largest leaps in defense technology developed in recent years – stealth airplanes and ships that avoid radar detection. These vehicles generally combine several technologies including the shape of the target’s surfaces to reflect energy away from the source and radar-absorbent materials. However, if a ship’s or aircraft’s antenna is to operate properly it cannot be completely covered. That makes it one of the remaining components with a large radar cross section (RCS) that can essentially destroy the overall system’s invisibility to radar.

The RCS depends on the polarization and frequency of the incident wave. When an electromagnetic wave is incident on a target, electric currents are induced in the target and a secondary radiation from that target produces a scattered wave.

The scattered field is partially reflected straight back to the source of the incident wave, and this is the principle upon which radar is based. The peak reflected wave is related to the standard antenna gain and its peak effective surface area. In this case, there is an ironic twist: Antenna designers normally look to maximize antenna gain, but to lower the RCS they must do the opposite of what they normally do, specifically reduce the gain.

One way around this problem is to employ a frequency selective surface (FSS). It consists of a pattern of shaped holes or surfaces on a substrate and essentially creates a bandpass filter. For instance, in the intended frequency range, where radio operators are transmitting or receiving, the antenna acts as normal. At other frequencies, the FSS absorbs rather than scatters incident radiation. Antennas are generally housed in a protective enclosure called a radome; for aircraft, often located at the nose. If that enclosure is made of such an FSS, its RCS is significantly reduced at all but the operating frequencies.
 

Geometric patterns as a filter

Frequency selective surfaces are usually constructed from periodically arranged metallic patterns of an arbitrary geometry. They have openings similar to patches within a metallic screen (see Figure 1 below). The performance of a FSS is linked to its shape, thickness, choice of substrate, and the phasing between individual elements. Altran engineers have focused on the physical configurations and the resonant frequencies for certain bandwidths, and COMSOL Multiphysics software has been an invaluable tool in these studies.

As shown in Figure 1, an FSS consists of a series of geometrical objects. The FSS can be electrically large in terms of wavelength with very many instances of the object, which would make simulating the entire surface extremely cumbersome and expensive in terms of compute power and time. COMSOL Multiphysics has a very convenient answer to this problem in the periodic boundary condition (PBC) feature. It allows the simulation of a single cell unit and thus a less time consuming process (see Figure 2 below). This feature provides for continuity of the electric and magnetic fields that generates equivalent results as if an entire array of objects had been simulated.

Researchers were very impressed with the time and memory savings possible with a PBC while keeping the level of accuracy needed to study the behavior of a given geometry. For a simple structure without a dielectric substrate, engineers estimate it cuts simulation time by a factor of 100; for a very large electrical structure, this could even be 1,000 times or more.

Figure 3 shows an example FSS made of simple metallic strips surrounded by air. The simulation mesh was created for one of the metallic strips, and as can be seen from the frequency plot (bottom left), it has a passband in the region of 40GHz.
 

To validate the simulation, Altran engineers first analyzed a case study already in the literature and replicated the known results in COMSOL Multiphysics with the aim of tuning the simulation procedure. In a second step, engineers took this same validated simulation and modeled other types of FSS while changing geometries and materials and evaluating the impact of these changes on FSS performance.

Altran researchers have used the software to investigate the frequency responses of a variety of simple shapes and sizes, and how they are distributed on a surface. It is also possible to make the design more complicated by using two structures with complementary behavior. In this way one can create a design with multiple resonant frequencies.

The ability to try any number of shapes highlights the capacity of the software to be efficient in finding a good solution. The alternative would be actually fabricating various shapes for the FSS and physically testing them, which would involve far more time and expense. With modeling, in a few minutes we can determine if a pattern is worth pursuing in detail.

Altran engineers are now starting to expand their model to include the effects of the dielectric substrate. In addition, they hope to start working with optimization algorithms soon to help in cases where they face constraints such as maximum unit cell size.

COMSOL Inc.
www.comsol.com

Altran Group
www.altran-na.com

About the author: Fabio Costa is a manager consultant at Altran and can be contacted at fabio.costa@altran.com.