EnginSoft - Conference Abstracts

EnginSoft International Conference 2009
CAE Technologies for Industry

Multi-Objective Optimization of Isoflux Antenna

Ravanelli Rodolfo - Thales Alenia Space
Iannicelli Caterina - Thales Alenia Space

Abstract

The Power Data Handling and Transmission (PDHT) antennas are basic payloads on low earth orbit (LEO) satellites. They are characterized by an antenna pattern shaped to compensate different gain attenuation relevant to satellite slant with respect to ground station (isoflux pattern). The challenging problem of the antenna design is the structure optimisation to meet more requirements as the gain definition on desired mask on extended coverage area (up to 64 degs in elevation and complete azimuth range), cross-polarisation discrimination, amplitude and phase ripples in band, return loss. The ThalesAleniaSpace first exemplary of a PDHT antenna was developed for RADARSAT and COSMO-SKY Med missions about ten years ago. The basic antenna architecture consists of a corrugated planar surface with cylindrical symmetry excited by quartz loaded launcher. The electromagnetic (e.m.) analysis of this geometry was possible by simple 2D Method of Moment modelling. The geometry configuration was optimized using simple minimization techniques based on quasi-Newton searching method. For recent satellite missions new and more stringent performances are requested for the antenna pattern, especially in terms of cross-polarisation discrimination and operative frequency bandwidth. This entails a redesign of the electromagnetic structure: to improve the electrical performance a series of slot across the corrugation walls have been inserted. The break on the geometrical symmetry requires now more complex 3D e.m. description with large increasing on computation time and memory resources.
The complete geometry definition should require over 100 independent variables, so that the possibility to optimize the antenna performance becomes very strong demanding. This represents a typical multivariable and multi-objective optimisation problem which has been resolved by modeFRONTIER approach using the powerful capabilities of the optimisation ambient. Individuation of the more sensitive variables has been an important step in the optimisation starting phase. A DOE (Design of Experiment) technique has been used to understand better the input-output relationship and to reduce the design space dimension (number of variables); then the multi-objective optimization has been carried out. The achieved final structure meets the overall requirements with a very simple and regular geometry with robust design. The reduction of the size has been showed as a new solution with improvement on the antenna satellite allocation. The modeFRONTIER capabilities consent to obtain different RF solutions, which permit to individuate the more simple ones with respect to the manufacturing aspects. The final design is currently employed for three satellite mission in the frame of European Space Agency GMES mission.