Non-probabilistic Model Based Robust Design And Integrated Structural Electromagnetic Optimization Of Antennas

According to whether the influence of uncertainties is considered,engineering optimization design can be divided into two categories.The first one is deterministic optimization in which the uncertainties are not considered.For some certain problems,the uncertain factors may lead to the variations of system performances and even result in performance failure.Therefore the uncertain factors should be taken into account in the design process.Robust design is one of the effective ways to handle uncertain information and has received widely attention.In this dissertation,we take the antenna structure system as the main application object and both of the deterministic optimization and robust optimization are involved.Several problems existing in the application of optimization theory into practical engineering are mainly investigated.The main works can be described as follows:1.A new robustness index is developed based on the convex model and the concept of feasible ellipsoid.The index can be used to measure the robustness of a design with multiple performance variation constraints.This index also has clear physical meaning and is easy to obtain.A robust optimization model is built with this robustness index.Compared with traditional robust design method,the solving of new model does not need inner loop and can significantly decrease the computational burden of the robust design.2.For robust design problems with both random and interval uncertainties,a robustness index with clear meaning is introduced.This index represents the possibility that the design satisfy the performance variation constraints.A new robust optimization model is established with the robustness index.On the one hand,this model can improve the efficiency by omitting the double-loop sampling of uncertain factors.On the other hand,the new model can ensure the satisfaction of objective robustness requirement.Several numerical examples are illustrated to demonstrate the applicability and efficiency of the proposed algorithm.3.The classical interval arithmetic is incapable to consider the dependence between the uncertainties and therefore it will lead to interval overestimation problems when it is applied into antenna electromagnetic analysis.To solve this problem,when the excitation amplitudes of the array elements are interval parameters,a new interval analysis method that is based on Taylor expansion is introduced to address this problem.The new technique can produce more accurate electromagnetic performance intervals and can avoid the complex computation of classical interval arithmetic.The numerical examples have demonstrated that the overestimation problem can be alleviated significantly.4.The back-up structure topology design problem of spaceborne reflector antennas with electromagnetic performance as objective is investigated.By combining the integrated structure-electromagnetic analysis technique with genetic algorithm based topology optimization method,the mathematical formulation of the back-up structure design is developed.Compared with traditional topology methods which aim to minimize the structure strain energy,the proposed method can produce solutions with better electromagnetic performances.5.For large phased array antenna structure with supporting adjusting mechanism,a new integrated mechanism design method is developed.The existing design techniques always treat the antenna plate as a rigid body and an acceptable accuracy within the whole adjusting space cannot be guaranteed because of the array deformation.To address this,a new integrated optimization model is proposed to design the locations of actuating links,which take both plate accuracy and adjusting mechanism accuracy into consideration.The numerical simulation results shows that we can ensure the plate accuracy in the whole adjusting space and the design can be readily used in engineering practice.