Antenna System Evaluation
Antennas Mounted on Large and Complex Structures
Consider an example where an evaluation of the V/UHF and HF communication antennas performance on an aircraft is required. Consider a case where the communication antennas of specific concern are:
- A V/UHF antenna mounted at the top of the aircraft fuselage (referred to hereafter as the Top-Fin antenna). The Top-Fin antenna operates at V/UHF (30 MHz - 420 MHz) frequency band.
- A V/UHF antenna mounted at the bottom of the aircraft fuselage (referred to hereafter as the Bottom-Fin antenna). The Bottom-Fin antenna operates at V/UHF (30 MHz - 420 MHz) frequency band.
- An HF loop antenna mounted at the bottom of the aircraft fuselage behind the Bottom-Fin antenna (referred to hereafter as the HF antenna). The HF antenna operates at the HF (2 MHz - 30 MHz) frequency band.
The Method of Moments solution technique as implemented by the Numerical Electromagnetics Code (version 2), NEC2, or its re-written version SuperNEC, is used in this study. Prior to a rigorous theoretical evaluation, using the Method of Moments technique, the validation of the aircraft numerical models is required. The aim of this study is to validate the aircraft numerical models. Validation of the numerical models is achieved by comparing the computed radiation patterns to radiation patterns which were measured on scaled models of the aircraft.
An example of the approach followed during such investigation is summarized below:
- Construction of a l/20th and a l/30th scale models of the aircraft for pattern measurements at frequencies which correspond to full scale frequencies between 15 MHz and 420 MHz.
- Measurements of the antenna radiation pattern in the principle planes (yaw, pitch and roll) are performed in a compact range anechoic chamber.
- Generation of wire grid numerical models of the aircraft, from available technical drawings, for the frequencies of interest and simulation of the structures to compute the radiation patterns associated with each of the communications antennas.
- Verification of the computed results by comparison to the measurements.
The evaluation of the performance of antennas on aircraft during the design phase of communications system is an interesting and difficult problem, which is certainly of international interest. The antennas employed in the VHF/UHF frequency ranges could have fairly simple monopole patterns when mounted on infinite ground planes. It is however evident that their performance changes dramatically when mounted on an aircraft. Both the aircraft geometry and the antenna position on the fuselage determine the ultimate performance of the antenna.
The radiation pattern performance of an antenna used for HF communications is very important in order to determine the ranges and quality of communications which will be achieved. Computer simulation of HF antenna performance in this case is essential, since the antenna and aircraft are strongly coupled and the antenna/aircraft combination acts as the dominant radiator. This is quite different from the situation at the higher (V/UHF) frequencies where the aircraft does "interfere" with the antenna radiation pattern, but not all parts of the aircraft carries large currents which in themselves are important sources of radiation. Simulation of radiation pattern characteristics at HF taking the complete aircraft into account is possible because the long wavelength allows the complete aircraft to be simulated.
Different approaches for assessing antenna performance on airframes have been employed in the past. These include full scale pattern measurement, scale model measurement, computer simulation and often nothing more than designer intuition. Full scale measurements are very often impractical, since alterations to the antenna positions cannot be performed without structural modifications which are a costly route to assess performances. Even in the case where antennas need only be investigated in existing positions the measurement of full antenna patterns with the aircraft in flight is difficult and not very accurate. Scale model measurements offer a lower cost solution, but still require considerable manpower and equipment resources for full three dimensional radiation pattern measurements. Such measurements also suffer from inaccuracies as a result of measurement uncertainties and scaling problems. Furthermore, alternative antenna positions are not easily accommodated on the model. Computer simulation of antenna patterns is an option which renders considerable flexibility to evaluate different antenna configurations and three dimensional patterns are easily generated. Computer simulation is limited by the computational resources (time and memory) and requires modeling skills. Any resulting numerical errors, however, are difficult to quantify if computer simulation is the only method used for performance evaluation.
The optimum solution approach which is currently available, therefore, is a combination of limited scale model measurements combined with computer simulation. This approach has the following advantages:
- Limited measurements are necessary to allow for computer models to be verified. These measurements are an invaluable part of the numerical model development process, since many approximations are normally required and without such bench-mark results these are very difficult to justify.
- The two complementary methods for obtaining results gives a measure of confidence to the overall study.
- In many instances one of the methods may be more suitable for assessment of specific aspects and/or frequency ranges. At higher frequencies the numerical solutions may be impossible to obtain using current computer technology and a lower frequencies measurements often become difficult due to scaling.
Computer Simulation of Antenna Performance
Two numerical techniques for simulating the performance of antennas on aircraft are plausible depending on the frequency range and size of the aircraft
- The Geometric Theory of Diffraction (GTD) method which is an asymptotic approximation valid for electrically large structures. This technique is often applied at the UHF range and above for the assessment of antenna performance. It is difficult to accommodate actual antennas coupled to the diffracting elements since current coupling is neglected in GTD theory. The technique is hence more suitable when the source is removed from the structure, such as in the case of radar cross section (RCS) studies or when investigating the performance of a reflector antenna where the source is separate from the reflector.
- The Method of Moments (MoM) which is formulated from a set of integral equations describing interactions between elements constituting a structure. These interactions are used in a matrix equation to find the currents on the structure. The solution technique in this case is theoretically accurate at all frequencies as long as modelling guide lines are observed. The limitation to this method, as implemented by NEC2, is not due to the breakdown of theoretical assumptions, but rather as a result of the increased computational requirements when electrically large structures are of interest. This problem is overcome by using the SuperNEC program.
The MoM solution technique as implemented by SuperNEC which is based on the Numerical Electromagnetics Code (version 2), NEC2, is used in this study. The NEC2 program has been developed by Lawrence Livermore Laboratories under contract from the Naval Oceanic Systems Centre (NOSC). Since its inception in the seventies the program has gained an international reputation and has become a de-facto standard for MoM antenna simulation. The Sparse Iterative Method (SIM), which is implemented by the SuperNEC program, provides a faster solution to the MoM matrix equations than does LU-decomposition with forward and back substitution. The SIM produces a solution with computational tune proportional to N2, as oppose to the N3 tune dependence associated with LU-decomposition. The SIM is implemented in an object oriented MoM program, SuperNEC, which is functionally equivalent to NEC2 and which can be executed in a distributed fashion, utilising any available computing facility which is connected to a LAN, each solving a portion of the matrix. This approach enables the simulation of much larger structures faster when compared to the implementation of NEC2 which where the matrix is solved by a single CPU.
A structure has to be described in terms of wire segments for the purposes of MoM simulation. Wire grids, constructed according to well established guide lines, are used to model solid surfaces.
The first problem confronting the engineer is the data entry of the structure in terms of wire segments describing the aircraft fuselage. The numerical models of the aircraft were constructed using our Structural Interpolation and Gridding software package, SIG, developed for this purpose. The SIG package generates a three-dimensional grid model from a set of user defined cross-sectional cuts at points of abrupt change along the three-dimensional structure.
The aircraft is, therefore, divided into cross sections obtained from technical drawings. These cross sections are then entered into a data file in a format which could be interpreted by the SIG package. The SIG package produces a NEC2 file of the numerical model which is based on the frequency of operation specified. An example of a wire grid geometry of a helicopter at 220 MHz is shown in figure 2.1. The Top-Fin and Bottom-Fin antennas were modeled as 250 mm monopoles. The HF antenna was modeled as half a loop 2845 mm long which is connected to the bottom of the tail via 350 mm wires. The Top-Fin and HF loop antennas are visible on the grid geometry.
Scale Model Measurements to Validate Computed Results
A l/30th and a l/20th scale models of the aircraft are constructed from technical drawings. The aircraft are constructed from UREAL foam which is covered with polymer sheets. The models are then copper plated to ensure high conductivity. The antennas are incorporated into the models by guiding coaxial cable to the antenna positions and protruding an appropriate scaled measure to form the monopole corresponding to the Top-Fin and Bottom-Fin antennas respectively, and to connect the HF loop antenna.
Selected two dimensional pattern measurements are then performed on the l/30th and l/20th scale models using a compact range suitable for the frequency range 2 GHz-18 GHz and a shielded anechoic chamber suitable for measurements in the frequency range 300 MHz-3 GHz. The compact range consists of a large anechoic chamber, calibrated feed antenna, offset parabolic reflector and a positioner with 3 axes of freedom. In the shielded anechoic chamber a 300 MHz-3 GHz Log periodic Dipole Array, LPDA, is used.
Examples of Models Verification Results at HF, VHF and UHF
Appreciating the electrical size and geometrical complexity of most aircraft, the aim is have the numerically predicted radiation patterns to compare well with the measured radiation patterns. The numerical models are validated by means of comparison between measured and computed radiation patterns. Using the validated numerical models, SuperNEC is used to obtain the 3-dimensional
radiation patterns of the antennas. The 3-dimensional radiation patterns are then used for statistical communication system performance evaluation in terms of probability of communications, POC, for different flight profiles.
Comparison between measured and computed results for a helicopter: (a) Bottom-Fin side roll plane at 420 MHz (b) Top-Fin azimuth plane at 320 MHz (c) Top-Fin pitch plane at 220 MHz (d) Bottom-Fin pitch plane at 118 MHz (e) Top-Fin side roll plane at 67 MHz (f) HF antenna azimuth plane at 15 MHz.