標題: | 多激式平面喇叭系統之模擬與最佳化 Simulation and Optimization of Multiply Excited Panel Speakers |
作者: | 劉博文 Bowen Liu 白明憲 Mingsian Bai 機械工程學系 |
關鍵字: | 平面喇叭;基因演算法;聲壓;聲功率;panel speakers;assumed -modes methods;genetic algorithms;sound pressure level;sound power density |
公開日期: | 2001 |
摘要: | 平面喇叭系統是由電力,機械和聲學三個子系統.所組成的。在本文中,我們考慮到空氣的聲學負載和平面喇叭結構上的耦合。並將致動器的等效阻抗,平板的機械阻抗和空氣的聲學阻抗結合,我們可以經由致動器的輸入電壓來換算平板上的表面速度。再藉由雷利積分(Rayleigh’s Integral),我們可以將平板的表面速度換算成空間中任一點的聲壓。在我們的模擬中包含了平面喇叭的頻率響應,聲功率和方向響應,並且比較了有無聲學負載的差異。為了減輕平面喇叭在高頻時的聲束效應(beaming effect)我們藉由基因演算法(genetic algorithms)來做最佳化的處理。在我們的最佳化中我們的設計變數是致動器的位置和信號的延遲。最後我們做了實驗來印證我們先前的模擬。發現模擬和實驗的結果有些誤差,主因在於模擬和實驗的邊界條件的差異。但在最佳化的實驗中,我們發現有較佳的全向性以及效率的提升,驗證了之前最佳化的結果。 The panel speaker consists of electrical, mechanical and acoustical sub-systems. We consider that the panel speaker system has experienced over the full interaction condition. In this paper, we construct the multiply excited panel loudspeaker system, and focus on the flexible panel with several exciters. Considering the whole system in impedance analogy circuit, we can derive the surface velocity of panel from the input voltage of exciter. From the Rayleigh’s integral, we can calculate the sound pressure in the space by the surface velocity of panel. Furthermore, our simulation includes frequency response, sound power, and directional response of the panel speaker. To reduce the beaming effect of panel speaker at high frequencies, the optimization of the panel loudspeaker in efficiency and omni-directionality is carried out by the genetic algorithms (GAs). The design variables include positions of exciters and the delay of input signal to exciters. The procedure of experimental design and measurement are carried out to prove our previous simulations. The result shows there is an error between simulation and experiment. The main reason is the error of boundary condition between simulation and experiment. However, in the experiment of optimization we discover that the panel has better efficiency and omni-directionality after optimization. It proves our previous optimization. |
URI: | http://140.113.39.130/cdrfb3/record/nctu/#NT900489010 http://hdl.handle.net/11536/69125 |
Appears in Collections: | Thesis |