【作者】 覃延明；

【导师】 廖成；

【作者基本信息】 西南交通大学 ， 电磁场与微波技术， 2009， 博士

【摘要】 超宽带雷达、超宽带通信、超宽带成像和超宽带电磁脉冲武器等领域的飞速发展,成为推动超宽带天线研究的巨大动力,并对超宽带天线提出了越来越高的要求。在众多形式的超宽带天线中,TEM(transverseelectromagnetic)喇叭天线及其变形天线是常用的高功率超宽带天线,工作于TEM模,定向辐射能力较强,易于工程实现,在超宽频带内具有恒阻抗特性,可以有效辐射时域窄脉冲信号。由于馈源、尺寸和功率等因素的限制,宜于采用阵列技术,以方便控制天线的辐射特性,减小波束宽度、降低副瓣电平和提高辐射功率,在诸如雷达系统、电子对抗、地下传感等众多军事和民用领域也有着非常广阔的应用前景。对超宽带天线阵列的分析存在一定难度,而对它的综合或优化将更具有挑战性,由于超宽带天线阵列中的优化变量与优化目标呈强烈的非线性关系,传统的优化方法往往无能为力--尽管如此,研究较为完善的适合于超宽带天线阵列优化的方法却是非常必要和迫切的,它将直接影响到超宽带天线阵列的各项性能指标以及阵列快速优化设计的实现。本文将以时域有限差分法(Finite-Difference Time-Domain Method,FDTD)为电磁辐射正问题的求解基础,以全局优化算法--微遗传算法、改进微遗传算法及其多目标进化算法分别处理天线及其阵列的优化问题(逆问题)。因此,论文首先对应用于超宽带TEM喇叭天线分析的数值方法--时域有限差分方法的特点和算法基本设置进行了简单介绍;对天线辐射问题的相关设置以及应用于超宽带TEM喇叭天线馈电的一维传输线馈电模型进行了描述;详细论述了获取天线远场信息的近-远场外推算法,对天线辐射问题应用改进的快速近-远场转换算法,对等效面近场数据在时域或空域上进行抽样,以尽量少的近场数据,实现不失真的远场转换,节约了计算时间,并给出了计算实例。论文接着介绍了描述超宽带天线时域特性的参数定义以及天线建模方法,并对其正确性进行了验证;详细论述了超宽带TEM喇叭天线的辐射特性。分析了各天线几何参数和激励源脉宽变化对其辐射特性的影响,介绍了超宽带天线阵列叠加原理,对叠状、槽状、面阵均匀排列和非均匀排列方式的阵列辐射特性进行分析,仿真分析了阵元间互耦影响和阵列波束电扫描特性;此外,还对超宽带TEM喇叭天线的变形结构--介质加载TEM喇叭天线和加脊TEM喇叭天线进行仿真模拟,并分析了天线的基本特性。随后重点论述了实现超宽带TEM喇叭天线优化设计所涉及的优化算法。介绍了单目标优化问题和多目标优化问题的不同处理方式,简述了传统优化算法的优缺点,随机优化方法特点和算法性能评估的相关准则;介绍了微遗传算法的实现过程,并对其交叉和变异操作等进行改进,提出了改进微遗传算法(Improve Micro Genetic Algorithm,I-MGA);针对超宽带天线多目标优化问题,基于Pareto最优解的概念,设计了基于自适应权重聚合和基于Pareto选择机制这两种不同形式的改进多目标微遗传算法;针对单目标和多目标优化问题的测试函数分别进行了性能测试,并与其它多种常用优化算法进行了比较。其后将微遗传算法及其改进算法与时域有限差分技术结合起来应用于超宽带天线阵列的优化问题。以天线阵列的增益、主瓣波束宽度、副瓣电平等作为优化目标,把难以求解的多阶方程组转化为适应度的约束条件,通过微遗传算法及其改进算法结合FDTD优化其电气、几何设计参数等,解决超宽带TEM喇叭天线阵的优化设计问题。论文论述了单天线优化及阵列优化的情况;对于多目标优化问题,分别以多目标加权法和基于Pareto的方法进行优化;为进一步提高超宽带TEM喇叭天线阵列的优化速度,提出了组合优化方法以缩小优化搜索范围,在不考虑阵元互耦情况下对天线阵列进行初次优化以获取先验知识,并根据初次优化的结果,在考虑阵元互耦情况下进行二次优化,以获取优化结果。最后,对论文工作进行了归纳总结并对下一步工作进行了展望。本文结合时域有限差分方法,从研究全时域的优化算法入手,完善了超宽带天线阵列的优化方法,从而实现超宽带天线TEM喇叭天线阵列的优化设计。更多还原

【Abstract】 The rapid development in the fields of ultra wide band (UWB) radar, UWB communications, UWB imaging and UWB electromagnetic pulse weapon and so on becomes a tremendous impetus for studying UWB antenna while ever-increasing requirements are proposed for UWB antenna design. Among various forms of UWB antennas, transverse electromagnetic (TEM) horn antenna and its transformative forms are commonly used in domain of high power microwave, working under TEM mode.They have strong-directional radiation and are easy to project implementing.They also have constant impedance characteristics in very broad frequency range and radiate time-domain narrow pulse signal effectively. Because of the limitation factors of feed, size, power, etc. and consideration of antenna-radiation controllability, it is fit for applying array technology to reduce the beam width and side lobe level and improve the radiation power. They have wide application prospects in the realm of military and civil, such as radar system, electronic warfare, underground detection and the like.There are not only some difficulties in UWB antenna analyzing, but also more challenges in UWB antenna synthesis and optimization. Because of the strong nonlinearity between the optimized variables and optimized objects of UWB antenna array, the traditional optimization methods sometimes can not deal with it. In spite of this, studying on more suitable and high-powered optimization methods for UWB antenna array is very necessary and urgent, which may influences the performance and optimization speed of antenna array directly.In this dissertation, finite-difference time-domain method (FDTD) is applied to be the basic analysis method for electromagnetic radiation problems and micro genetic algorithm (MGA), improved micro genetic algorithm (I-MGA) and improved multi-objective evolutiaonary micro-genetic algorithms are used to process the optimization problems of antenna and array respectively.Firstly, FDTD is introduced here as UWB antennas analysis method, whose advantages and the involved algorithms will be discussed. A simple FDTD model for transient excitation by transmission lines is expound, which divides the computational region into excitation source region and antenna region to reduce computational cost. The improved rapid time-domain near-far extrapolation method based on FDTD calculations is used to obtain far-field information of antennas. In order to minimize the calculation data for far-field transforming, the data in Huygens’ equivalent surface are sampled in space-domain or time-domain, which saves the computing time consumedly. An example is given.Secondly, the definition of parameters that describe the time-domain characteristics of UWB antenna and the method for antenna modeling are introduced. The validity of program is proved subsequently. After that the radiation characteristics of TEM horn antenna are told out. The impact of antenna variables’ changing on its radiation characteristic is analyzed. Then the superposition principle of UWB antenna array, the different influence of array arrangement style, mutual coupling effect between array elements and time-domain electric scanning characteristic are introduced. The transformative antennas of TEM horn antenna, dielectric loaded TEM antenna and double-ridged TEM horn antenna, are also simulated, whose characteristics are analyzed as well.Thirdly, the involved optimization algorithms focusing on the realization of UWB TEM horn antenna design are mainly discussed here. The different process measures for sigle-objective and multi-objective optimization problem will be mentioned.The advantage and disadvantage of traditional optimization algorithms and the characteristics of stochastic optimization algorithms are introduced subsequently.The criterion of performance evaluation of the stochastic optimization algorithms will be given as well. Micro genetic algorithm (MGA) is compared with genetic algorithm (GA). The former who searches more efficiently is improved here and the improved micro genetic algorithm (I-MGA) is proposed, whose performance is tested. In addition, two different forms of improved multi-objective MGA are designed based on non-domain Pareto optimal concept for multi-objective optimization problem.The two kinds of multi-objective optimization algirithms are tested by testing functions, compared with the other optimization algorithm. The detail is presented.In addition, MGA, I-MGA and improved multi-objective evolutiaonary micro-genetic algorithms are combined with FDTD respectively for UWB antenna array optimization problem. We take gain, beam-width, side lobe level, etc. as the optimization objects, changeing the high-order equation groups that are hard to solve to be the restrict conditions of evaluation fitness. MGA and I-MGA are applied to optimize the electric and geometric design variables and so on while radiation problem is analyzing by FDTD. This dissertation discusses the single antenna optimization and array optimization problems here. For multi-objective optimization problem, we use weighted sum approach and Pareto-based approach respectively to proceed. In order to further accelerate the optimization speed, combination-optimization method is proposed. Regardless of couple effects of array elements, we optimize the antenna array for the first time to acquire prior knowledge. After that, according to the first optimization results, we take the couple effects into account and do secondary optimization to get the final outcome.Finally, the work is summarized and the next research directions are predicted.In this dissertation, we start from full-time domain optimization algorithm and improve the performent of the optimization method of UWB antenna array. Combined with FDTD, the optimization of UWB TEM antenna array is achieved.更多还原