Optimizing Processing Time of Radio-Astronomy Antenna Simulations Using FEKO

Abstract

─ The far-field pattern of a geometrically large and complex antenna used in low-frequency radio astronomy is computationally expensive to simulate on electromagnetic simulators, such as FEKO. For example, one station of the Square Kilometer Array, which consists of 256 log-periodic antenna elements, will take years to simulate using the full CAD model for the full operational frequency band. This paper focuses on reducing the simulation time for a single antenna element by simplifying the simulation model, thus decreasing the number of unknowns that have to be solved in a simulation. An iterative process for optimizing the simplification of such an element is described, while keeping the reflection coefficient within 1 dB absolute mean deviation of the measured data. After four iterations, the amount of unknowns to be solved, which includes the number of triangles and segments, was reduced from 29,307 to 11,991. This decreased the computation time by 86.5%, making array simulations feasible. Using the techniques described in the paper, other antenna constructions can benefit from it and be simulated more efficiently. Index Terms ─ Antenna, FEKO, optimization techniques, radio astronomy, square-kilometer array