本論文之目的在求取使複材三明治樑之彎曲剛性有最佳表現之因子水準組合,使用田口實驗法進行實驗規劃,依照美國標準測試ASTM C393作為三明治樑測試之依據,紀錄試片之力對位移關係,計算出樑構件之彎曲剛性值。將所得數據進行訊號雜訊比及變異數分析來探討各因子對彎曲剛性之影響趨勢,分析出最佳之因子組合,之後進行驗證實驗來確認最佳化之複材三明治樑。 文中所探討的對象為纖維布搭配泡棉芯材的三明治組合,所探討變因為纖維板種類,泡棉種類以及黏膠種類,又各取三個水準值進行探討。配合所選定之水準數後,選擇L9(34)直交表來設計配置9組因子水準組合之測試試件,測得力對位移關係,求取彎曲剛性,其目標特性為望大的,依照望大特性之定義計算S/N比及ANOVA分析,發現在碳纖維疊層、密度100kg/m3之泡棉及Kit 138黏膠三因子水準組合下,有最佳化之剛性,之之後進行驗證實驗,也確認此預測值之準確性。 本論文應用田口法探討樑之各項參數對於彎曲剛性之影響程度。可以以最少之實驗次數來求得各項參數之影響,達到瞭解三明治樑之彎曲性能的效果。在不降低結果可靠度的前提下,可以有效節省實驗期程及成本,故此設計方法可以作為其他實驗規劃設計之參考。
The purpose of present study was to find out a best combination level of factors, which would give an optimal bending stiffness of a composite sandwich structure. Taguchi method was used for experiment planning, while bending test was applied based on a standard testing procedure, ASTM-C393, the force-displacement relationship was output. Output data were then used to do S/N ratio and ANOVA analysis, to investigate the contribution of each factor to bending stiffness, and the optimal combination of levels could be predicted. Verification experiment was then carried out to confirm the optimized sandwich beam. Sandwich beams made of composite laminate with foam core were brought to discussion, and the factors considered were: type of laminate, core, and adhesive, and each factor with three levels. After levels of factor were chosen, L9 orthogonal array was used to set each level of factor to form 9 sets of testament. Since the property of the objective function, bending stiffness, which is the larger-the-better, S/N ratio could be calculated, and AVOVA analysis was carried out. The results showed with the combination of CFRP lamination, and foam with density of 100kg/m3, and adhesive Kit 138, there would be an optimal stiffness. Verification test was then carried out to verify the accuracy of the predict value. With the application of Taguchi method, the contributions of each factor to objective function were investigated. Minimal numbers of tests were required to get the result. Without reducing accuracy, the cost and time could be effectively saved. The method was recommended to be a great tool to experiment design.