Error Control via Tension for an Array of Flexible Square Antenna Panels

Abstract

The feasibility of combined global and component level error control exercised singularly by the global prestress which maintains the integrity of a tension structure is explored via a design study for a flat tensioned phased array antenna. Considered, in particular, is a 2×5 grid of square antenna panels, integrated with a network of cable links connecting the panels to four catenaries along the array edges, externally supported at four corners. It is implied that overall flatness is achiev ed via the planarity of support: by keeping one of the four suspension points on the plane subtended by the other three. Investigated in this specific context is the mitigation via the suspension tension itself of environ mental perturbations to both global and component flatness. Namely, how well the tension can suppress out-of-fl atness due to dynamics (slew) and to panel thermal warping. The design is found to be clearly governed by panel thermal deformations. For the array of ten 1 m 2 panels of 5 kg mass each, if 1 o /s rotational acceleration is achieved with a 10 s slew as specified, dynamic out-of-flatness is comfortably reduced to below the 1 mm limit by tension as low as 40 N per panel edge (per panel row or colum). Order of magnitude greater tension is necessary to sufficiently suppress thermal deformations from a �T =15 o C gradient through the 32 mil panel thickness, as revealed with a finite element (FE) study. The FE study also provides insight into the intricate mechanics of the tension smoothing of the involved thermal deformations. In particular, the panels are found to deform under the thermal gradient into a diagonally twisted antisymmetric shape. Subsequent tension firs t removes the shape antisymmetry and concurrently reduces figure errors with a factor of roughly 2 to 7, dependin g on how it is applied. Further load continues error reduction with a lower rate to a certain limit. This limit is zero and is approached asymptotically if the tension is distributed continuously around the panel perimeter, but it is a finite error level otherwise. For the practical design, panel stretching with two types of cable arrangement is examined. One is a pair, the other is a triplet of cables, identical on all edges aroun d the panel. With cable pairs, error reduction is best when the cables are located at around 20% of the sides from the corners. The out-of-flatness of concern, specified as the deformed panel shape absolute depth, is redu ced below the required 1 mm threshold when each of the eight cables around the panel is pulled with about 530 N (120 lbf). This is more than ten times the 40 N tension per panel side that has been found to comfortably suppress transient deformations. The use of a third cable at the edge centers significantly improves perfo rmance. If this cable bears about 40% of the total edge load and the other two cables are next to the corners, the total tension sufficient to reduce out-of-flatness to 1 mm is only about 140 N (32 lbf). Two alternatives of how the catenaries are supported are also examined: direct attachment to fixed exterior locations, and support via short cable links inserted between an exterior support and the associated catenary endpoints. The design with no corner links performs slightly better, and it also offers some practical advantages.