傳輸線最佳化設計於微波被動電路之應用

Abstract

此篇論文結合基因演算法與傳輸線分析達成有效率的微波被動電路最佳化設計。傳統設計方法不是侷限於單頻設計就是窄頻設計。對於寬頻或雙頻之設計有其困難。為了解決這個問題，我們結合基因演算法及傳輸線分析方法來做最佳化設計。 在第一個應用例子裡，我們提出有效率的兩段式最佳化設計方法於w頻段高通波導管率波器之應用。此最佳化設計結合混合型基因演算法、傳輸線模型及全波分析。基於此方法，我們製作出19個量測結果良好的精小型高通波導率波器。 第二個例子使用基因演算法設計非均勻分佈線性陣列天線。此設計適於GSM/DCS/WCDMA基地台之應用。在此，我們提出基因演算法設計三頻平行饋入網路及Wilkinson功率分配器的方法。非均勻天線陣列的間距可由基因演算法調整，以避免雙主波瓣發生。模擬結果顯示非均勻天線陣列藉由平行饋入網路或Wilkinson功率分配器有良好的效能。此方法可完成三頻天線陣列之設計。 最後例子的應用是使用基因演算法設計雙頻率波器。第一種是利用雙頻元件，亦即雙頻共振器與雙頻反向器，來建構雙頻率波器。在此，我們使用複合式基因演算法尋找雙頻元件的組合及尺寸。此方法可能能尋求新的物理架構及探索出雙頻濾波器的物理機制。

This dissertation presents efficient optimization design based on genetic algorithms and transmission-line analysis approach for microwave passive circuit applications. Conventional design approaches of microwave passive components are constrained by either single-band or narrow-band operations. It encounters some difficulties in the design of components with wide-band or dual-band responses. To overcome this problem, one straightforward approach is to resort to optimization methods. Transmission-line based analysis approaches are especially efficient in optimization design. In the first application case, we propose an efficient two-phase optimization approach for a compact W-band double-plane stepped rectangular waveguide filter design. The optimization design combines genetic algorithms with the simplified transmission-line model and full-wave analysis. Numerical results show that the resultant waveguide filter design with 4 sections (total length 19.6 mm) is sufficient to meet the design goal and provides comparable performance to that with 8 sections (total length 35.6 mm) by the Chebyshev synthesis approach. Based on the present approach, nineteen compact high-pass waveguide filters have been implemented and measured at the W-band with satisfactory performance. The second case shows the methodology of non-uniformly spaced linear array design for GSM/DCS/WCDMA base station application using genetic algorithm. A new approach to facilitate the parallel corporate feeding network and the Wilkinson power divider for tri-band antenna array using genetic algorithm are proposed. The element spacing of non-uniformly spaced linear array is adjusted by genetic algorithm to avoid double main-beam problem. The simulation results show the non-uniformly spaced linear arrays fed by an 8-way Wilkinson power divider/Wilkinson power divider are suitable for a tri-band antenna array design. The last case is an application of dual-band filter design using genetic algorithms. All candidate elements, i.e. the resonators and the J-inverters of the dual pass-band filter, in the optimization process should inherently possess the arbitrary dual-band behaviors. The genetic algorithm searches for both the structure combination and dimensions of dual-band elements. Three filters are tested to validate this design in this work. The bandwidths of the first filter are 40% and 18.1% with center frequency 1.0 and 2.2 GHz, respectively. In addition, the bandwidths of the second filter are 18.18% and 25.06% with center frequency 0.88 and 1.94 GHz, respectively. The bandwidths of the third filter are 40% and 40% with center frequency 1.0 and 2.5 GHz, respectively. All the results of these three filters show satisfactory performances.