Research On A Large Deployable Antenna

In recent years, the large-aperture deployable antenna is desperately needed torealize long distance communication due to the increasing space exploration activitiesof human beings. In the thesis, inspired by the AstroMesh deployable antenna and thelarge deployable reflector, one kind of large deployable antenna (LDA) structure thathas lighter weight but a little lower fundamental frequency than the large deployablereflector structure is proposed to be applied to large-aperture deployable antennas. Inorder to avoid vibration caused by fast deployment speed of the LDA structure, abraking system is used to control the spring-actuated system. Moreover, the LDAstructure is designed to form a spherical surface found by the least square fittingmethod so that it can be symmetrical. Modeling method of the structure is based ongeometric method of coordinate transformation and two principles. In addition, thefitting errors of the terminal points are considered and deduced. After given a set ofknown parameters, the experiment model of the LDA structure is built and the fittingerrors are solved.Structural analysis of the deployable unit in the LDA structure is processed todetermine the geometric relationship among length of each beam, which realized thedeployment simultaneity of all beams. The nonlinear finite element method is used tomodel the structure in MATLAB, including numbering the structure before modeling.The fundamental frequency of the structure solved in MATLAB is very close to thatbuilt in ANSYS, which proves the finite element algorithmic routine in MATLAB iscorrect, which will be beneficial to following structure optimization.The composition of the cable network in the antenna is presented and analyzed,furthermore, an introduction to the force density method used to find form is offered.The cable network is found by the method combining the force density method andthe parabolic surface constraint. After analysis of the form-finding results of the frontand rear cable net, it is found that the pattern for the front and rear cable net must bethe same to form a parabolic surface when finding the shape of the rear cable net. Inaddition, the surface errors analysis of different patterns used for the cable net indicatethat triangular patterns have the minimum error among various patterns.Finally, a brief introduction to the single-object genetic algorithm and themultiple-object genetic algorithm is presented. The former is used to solve a part oflengths of beams in the deployable unit, and ensure that each cable in the net reachesa desired tension so as to improve the stability of the cable net after full expansion.The latter is utilized to optimize lengths and the cross-sectional dimension of beams and it is notice that the pareto optimal solution set is getting better with the increase ofiterations.