This application claims priority to European patent application 07446005.6 filed 20 Apr. 2007.
The present invention relates to the field of low signature antennas integrated in a vehicle structure.
There is a need today for creating a low radar signature for different objects such as e.g. aircrafts, i.e. to design aircrafts having a low radar visibility. Significant progress has been achieved in a number of problem areas as e.g.:
A number of solutions have been proposed for antennas with a low radar signature or a low Radar Cross Section, RCS.
There are two main problems with existing solutions for creating low RCS with low frequency antenna arrays integrated in a vehicle structure such as a wing edge. Henceforth a vehicle structure is exemplified by a wing edge. Firstly the elements in the antenna array need to be fairly large in order to be resonant, leading to large separations between antenna elements in the array and many grating lobes at higher frequencies. Grating lobes appear in antenna arrays with a periodic repetition of antenna elements and when the distance between elements in the array is greater than a half wavelength. At a frequency of 1 GHz (Giga Hertz) this critical distance is 15 cm.
Secondly the RCS of a straight cylindrical surface is proportional to the local radius of curvature of the surface and to the square of the length divided by the wavelength. Hence the RCS of a wing edge typically increases with frequency. For aero-dynamical reasons the radius of curvature needs to be fairly large with a high RCS as a result, especially at higher frequencies.
In order to reduce the RCS of metallic structures, e.g. including antenna elements, they are coated with Radar Absorbing Materials (RAM). Radar Absorbing Materials are characterized by complex permittivity and permeability values that usually vary with frequency. For planar stratified media with several layers with different properties there is a reflection for each transition and an attenuation of the wave inside the media. Using nonmagnetic purely dielectric media, both the reflections and the attenuation is increased with increasing imaginary part of the dielectric constant, hence there is a trade-off between high attenuation, ensuring low reflection from the inner metallic interface and low reflection from the outer interface. If the reduction in RCS is desired in a narrow frequency band, the thickness of a RAM-layer can be chosen in such way, that the attenuated reflection from the metallic surface has the same magnitude but opposite phase compared to the primary reflection, thereby cancelling it out. For wider frequency bands, this is not possible to accomplish but both the primary reflection and the secondary attenuated reflection need to be low. By using several layers with small change in dielectric properties, the reflection from each interface can be maintained low, while the attenuation is gradually increased, thereby reducing the total required thickness compared with the case when using a single layer with low permittivity material. Another way to accomplish low reflection in the first interface is to use a material with magnetic properties as well. However, the frequency behaviour of the permeability must match the frequency behaviour of the permittivity, and the reflections will only be low at incidence angles close to normal if the permittivity and permeability values are high.
Commercial RAM materials are generally designed to give a good RCS reduction performance in a wide frequency band and have a slow transition from low attenuation and high reflection at low frequencies to high attenuation and low reflection at high frequencies. When using this kind of material in the intended application, either the antenna losses will be unacceptably high or the RCS at medium range frequency will be too high.
Investigations have shown that it is possible to reduce the RCS levels over a frequency band in a threat sector in elevation by optimization of the material parameters and preferably also the shape of the inner profile of a RAM coated wing edge.
Hence there is a need to provide a method for manufacturing an antenna or antenna array and an antenna or antenna array with a low RCS value integrated in a structure having a large radius of curvature and at the same time accomplish a low attenuation of electromagnetic energy at low frequencies and a low reflection for incident waves at higher frequencies.
It is therefore the object of the invention to provide a method for manufacturing an antenna or antenna array, with an operating frequency band, comprising antenna elements integrated in a vehicle structure as well as an antenna or antenna array manufactured according to the method to solve the problem to achieve an antenna or antenna array with low RCS while at the same time accomplishing a low attenuation of electromagnetic energy at low frequencies and a low reflection for incident waves at higher frequencies.
This object is achieved by a method wherein a RAM structure, conforming to the shape of the vehicle structure and comprising at least one layer of RAM material with an inner surface facing the antenna element and an outer surface being an outer surface of the vehicle structure, is mounted in front of the antenna elements, each RAM-layer denoted i being defined by a thickness di and frequency dependent RAM properties:
relative permittivity ∈i,
relative permeability μi,
the frequency dependency of the RAM properties being tailored and the thickness di and the number of RAM layers being selected such that the RAM structure is substantially transparent in the operating band, reaching a predetermined Farfield pattern requirement, and simultaneously is an effective absorber, reaching a predetermined Radar Cross Section (RCS) requirement RCSth, at frequencies in a threat band comprising frequencies above the operating frequency band of the antenna, and an RCS requirement RCSop in the operating frequency band. The object is also achieved by an antenna or antenna array manufactured according to the method.
Normally the antenna or antenna array has a continuous operating frequency band, but the frequency band can also, within the scope of the invention, be divided in a number of bands, e.g. separate transmit and receive bands.
In prior art only a single RAM-layer with constant permittivity and permeability and also only incidence in the plane transverse to the wing axis has been considered. Although the wave is scattered in a cone away from the transmitter from an infinite long cylindrical structure for other incidence angles, the finite extent of the wing will introduce side-lobes pointing in the direction of incidence. These side-lobes will be proportional to the specular reflection in the elevation plane, why this reflection has to be considered as well. This is illustrated in
An advantage with the invention is that by tailoring the permittivity ∈ in the RAM layers it will be possible to obtain a faster transition from low attenuation and high reflection at low frequencies to high attenuation and low reflection at high frequencies. This is illustrated in the diagram of
The invention can advantageously be implemented on wing edges and an outer protective layer can be applied to the RAM structure to increase the mechanical strength of the RAM structure.
The invention can be applied on several types of antenna elements (dipoles, crossed dipoles, patches, fragmented patches etc). It is also possible to apply the invention using different feeds (slots, probes, balanced, unbalanced, etc).
a schematically shows incident and specular waves
b schematically shows RCS from side lobes of incident waves
The invention will in the following be described in detail with reference to the drawings.
A cross section of an upper half of a wing structure 501 with a RAM structure 502, having an inner surface 508 and outer surface 509, is shown in
Each layer i in a multilayered RAM is described by their material properties; relative permittivity ∈i, relative permeability μi and layer thickness di. The tangential component of the propagation vector for a plane wave travelling with angle θ from the normal in vacuum is k0 sin θ in all layers, where
is the wave number in vacuum.
For each interface, the tangential components of both the E-field and H-field are continuous; leading to that the incident wave is split into a transmitted wave and a reflected wave, travelling the opposite normal direction as the incident wave.
The normal component of the propagation vector in layer i is k0√{square root over (∈iμi−sin2 θ)}, since the tangential component is the same in each layer. The H-field is perpendicular to the E-field and the direction of propagation, and the E-field is perpendicular to the direction of propagation. The amplitude of the E-field is
times, η0=the characteristic impedance in free space, the amplitude of the H-field, hence the tangential component of the E-field is
times the tangential component of the H-field, when the E-field is in the plane of incidence.
When the E-field is perpendicular to the plane of incidence, the tangential component of the E-field is
times the tangential component of the H-field. For other polarisations, the incident wave can be decomposed into a component in the plane of incidence (parallel or TM polarization) and a component perpendicular to the plane of incidence (perpendicular or TE polarization), which can be treated separately.
When the incident wave meets the upper interface, one part of the wave energy is transmitted through the interface and the rest is reflected in the so called specular direction. The amplitude of the reflected wave is determined by that the tangential components of both the H-field and E-field are continuous, giving the relation:
where
for TE polarization and
for TM polarization. The amplitude of the transmitted wave is given by
and this wave is propagated and attenuated before it reaches the next interface.
Eref=reflected E-field
Einc=incident E-field
Etrans=transmitted E-field towards next layer.
Zi=impedance of layer i
For high frequencies the attenuation of the wave is so high, that it does not reach the next interface, the primary reflection is then dominant and should be kept as low as possible. One way of doing this, is to use a material with μi=∈i, making the reflection coefficient zero at normal incidence. One drawback with this approach is that the reflection coefficient increase rapidly with increasing incidence angles, if the magnitude of μi=∈i is large. Further, both the permittivity and the permeability are functions of frequency, and it might be difficult to match those over a large frequency band.
A commonly used model for describing the frequency dependency of the relative dielectric constant ∈r, or permittivity, is the Lorentz model, having 5 parameters according to:
where ∈∞ is the high frequency limit, ∈s the value at zero frequency, frel the relaxation frequency, f0 the resonance frequency, ∈0 the value in vacuum and finally σe the conductivity at zero frequency. Letting the resonance frequency approach infinity reduces the model to the Debye model with 4 parameters:
As an example consider a mixture of two materials, one base material with low dielectric constant close to 1 for all frequencies and the other with ∈∞=1, frel=4 GHz and f0=8 GHz independently of inclusion material volume fraction and where the other parameters, as ∈s and σe, are a function of the volume fraction according to the Maxwell Garnett mixing formula which is the simplest and most widely used model for description of composite media at comparatively low concentrations of inclusions. By proper choice of the volume fraction, values according to
In practice, materials with such low dielectric constant as in the outer layer in the example above have poor mechanical properties. In this example the arrangement has to be protected with a thin layer of mechanical stability, often having a larger dielectric constant or permittivity. The material properties of this layer have to be taken into account in the optimization of the structure.
As a comparison with what is typically achieved with commercial RAMs, data from a user supplied data sheet is fitted to a Debye model. The data was only available between 5 and 18 GHz and the original data is displayed with solid curves, the fitted data is shown with dashed curves in
When
The curve shape of the RAM-layers can be calculated using the Continuum Sensitivity Based approach for optimization. This is done by solving the E-field for TM polarization or the H-field for TE polarization for a set of frequencies, incidence angles and parameter values. The character σ is conventionally used for denoting RCS. Henceforth σ is therefore used for RCS and should not be mixed up with σe used for conductivity. The change ∂σ of the radar cross section by a small displacement ∂ξi in the normal direction of an interface between two different media i and i+1 can be expressed as an integral over the interface of an expression involving the solution to the problem and the solution of the adjoint problem (as described by Yongtao Yang in “Continuum Sensitivity Based Shape and Material Optimization for Microwave Applications”, Ch almers University of Technology, 2006, ISBN 91-7291-73-7):
for TM polarisation and
for TE polarisation. Similarly, the change of RCS by a small change ∂∈i and ∂μi in material parameters is given by the surface integrals
The RCS value is calculated according to:
∈i=relative permittivity
μi=relative permeability
k0=wave number in vacuum
∫Γ=line integral at interface between media i+1 and i
∫S
|E0|2=the square of the incident E-field amplitude
|H0|2=the square of the incident H-field amplitude
∇E=the gradient of the E-field
∇Ea=the gradient of the adjoint E-field as defined by Yongtao Yang in “Continuum Sensitivity Based Shape and Material Optimization for Microwave Applications”
∇H=the gradient of the H-field
∇Ha=the gradient of the adjoint H-field as defined by Yongtao Yang in “Continuum Sensitivity Based Shape and Material Optimization for Microwave Applications”
|Es|2=the square of the scattered E-field amplitude at distance R
R=distance from scattering source
The formulas for the RCS value and gradients above are valid for calculations in 2D but when necessary, calculations can also be performed in 3D using corresponding 3D formulas.
Also the H-field at any point on the inner PEC interface can be determined for each set of values. By reciprocity, the far field radiation pattern of a magnetic current line source placed in the corresponding point can be determined. The radiation efficiency can be determined by integrating the Farfield radiation pattern and the power delivered into the media surrounding the line source. The Farfield radiation pattern is defined as the vector product between the E- and H-field. All calculations of the Farfield in this description are made for both TE and TM polarization. In a corresponding way the E-field at any point on the inner PEC interface can be determined and by reciprocity the far field radiation pattern of an electric current line source placed in the corresponding point can be determined.
A suitable cost-function involving RCS, desired radiation pattern and efficiency has to be minimized, the partial derivatives of the cost function with respect to the design parameters can be determined by the chain rule, leading to fast convergence of gradient search algorithms.
Investigating the responses shown in
The method for the invention shall now be described with reference to the flow chart in
In 1203 an RCSop value (RCS in operating band) for cross-polarized waves with a frequency in the operating band is calculated for the selected initial shape assuming one RAM layer with ∈i=1, i.e. for air, according to formula:
RCSop gradients are also calculated according to:
for TM polarization and
for TE polarization
in order to decide whether a minimum RCSop value has been obtained for the selected parameter set. The calculations are made both for TE (Transverse Electric) and TM (Transverse Magnetic) polarizations.
In 1204 the calculated RCSop value is compared to the predetermined RCSop requirement for the operating band with one RAM-layer and ∈i=1.
If the requirement is not met the initial shape is updated with a new parameter set in 1205 and new calculations are made according to 1203. The resulted RCS value is again compared with predetermined requirements and if the requirement is met the procedure continuous to 1206, otherwise a new loop is made through 1205 and 1203 until the requirement is met.
In 1206 the Farfield in the operating band is calculated with ∈i=1 and with an initial position 1207 of the antenna elements along the initial shape with the tangential points 511 and 512 of the inner surface 508 mounted to the antenna element surface. The Farfield is calculated using a CEM (Computational Electro Magnetic) simulation with a magnetic or electric current line source at the position of the antenna element.
The calculations are made both for TE (Transverse Electric) and TM (Transverse Magnetic) polarizations. In 1208 a comparison is made with predetermined values for the Farfield. If requirements are not met positions of the antenna elements are updated in 1209 and new calculations are made according to 1206. A new comparison with predetermined requirements is made in 1208 and if the requirement is met the procedure continuous to 1211, otherwise a new loop is made through 1209 and 1206 until the requirement is met.
In 1210 a one layer RAM is selected with an er-value calculated according to the Debye model:
where ∈r=relative permittivity for the RAM-layer, ∈s=relative permittivity for the RAM-layer at zero frequency, ∈∞=relative permittivity for the RAM-layer at high frequency limit, ∈0=relative permittivity for the RAM-layer at a resonance frequency of the RAM-material, f=operating frequency of the antenna, frel=relaxation frequency, σe=conductivity at zero frequency. Examples of how to achieve different ∈r-values have been described above.
In 1211 following calculations are now made with the selected shape of the inner surface, antenna element positions and ∈r-value:
A comparison is made in 1212 against predetermined requirements for the Farfield in operating band and the RCSth values in the threat band for both TE and TM polarizations. If the requirements are met the design is finalized in 1213 and if not, a check is made in 1214 to see if a minimum is reached for a cost function including the Farfield pattern and the RCSth value. A cost function is an optimization algorithm which reaches a minimum when the parameters are optimized according to the conditions in the algorithm as further described above. If a cost function minimum is not reached the material parameter set made in 1210 is updated in 1215 and new calculations are made in 1211. A new comparison is made in 1212, if OK the design is finalized, otherwise a new check in 1214 is made. The loop continues until the procedure ends up in 1213 or when it is established in 1214 that the cost function minima is obtained. The procedure then continues to 1216 where the number of RAM-layers is increased by one and additional material parameters as e.g. interface shape parameters and thicknesses of RAM-layers are introduced. New calculations are then made in 1211 and the loop continues until the requirements are met in 1212 and the design is finalized.
Normally the calculation are made for the relative permeability μi=1. However the scope of the invention is not limited to a fixed μi-value, but this value can also be used as a variable parameter in the design process.
The invention is not limited to the embodiments above, but may vary freely within the scope of the appended claims.
Number | Date | Country | Kind |
---|---|---|---|
07446005.6 | Apr 2007 | EP | regional |