以臺灣櫸作為柳杉結構材之補強材料探討

本研究利用臺灣櫸（Zelkova serrata）做為柳杉（Cryptomeria japonica）實尺寸結構用材之補強材料，以目視分等、應力波法及抗彎彈性模數試驗進行膠合單元之評估，計算出膠合樑之抗彎彈性模數預測值（Ebp），配置後膠合成膠合樑進行抗彎強度性質、膠合剪斷性質、剝離性質及縱向壓縮強度性質之分析與評估。利用測量出之膠合單元抗彎彈性模數（MOE），進行強度配置後，以彎曲剛性（EI）相等之原則，計算以各層膠合單元配置過後膠合樑之Ebp，與抗彎彈性模數實驗值（Ebo）做相關性分析，其R2值為0.74，具有顯著正相關，代表預測值具有參考價值，可達到預測效果。抗彎試驗結果顯示，若只考慮抗彎強度（MOR）時，以17 mm臺灣櫸補強於柳杉上側、以17 mm臺灣櫸補強於柳杉上下兩側及以23 mm臺灣櫸補強於柳杉上下兩側皆有顯著之強度，而以17 mm臺灣櫸補強於柳杉上側之試材，臺灣櫸體積占比最少，因此有最佳的成本效益，其Ebo為9.7 GPa，相較柳杉對照組提升了14.5%，MOR值為74.8 MPa，相較柳杉對照組提升了49.8%。而若只考慮Ebo時，僅有以17 mm臺灣櫸補強於柳杉上下兩側及以23 mm臺灣櫸補強於柳杉上下兩側獲得顯著之Ebo值，因此僅需使用以17 mm臺灣櫸補強於柳杉上下兩側就能獲得顯著之Ebo與MOR值，Ebo值為14.6 GPa，較柳杉對照組提升了29.2%，MOR值為79.3 MPa，較柳杉對照組提升了58.8%。另外，實大樑及小尺寸之相同比例試材，比較Ebo時，與本研究有相似之結果，但比較MOR時，實大梁以17 mm臺灣櫸補強於上方之試材有顯著強度，小尺寸試材則無差異。膠合樑膠合性質方面，以CNS 11031標準檢驗，只有30.2%符合膠合剪斷標準，在煮沸剝離試驗中，本次膠合樑只適用 Ⅱ 及 Ⅲ 類使用環境。根據上述膠合剪斷試驗及煮沸剝離試驗結果，可發現本試驗膠合層之膠合性質不佳，推測原因與製程中單面佈膠、加壓不均及墊木過薄有關，也與材料本身收縮率差異過大以及臺灣櫸表面光滑具油脂有關。因此如要改善此現象，除了改善製程，應考慮將臺灣櫸先進行蒸煮製程，以增加表面性質，或者需替換其他尺寸安定性相似之材料。於縱向壓縮試驗中，各組膠合樑之抗壓強度是無顯著差異的，而破壞模式依照ASTM D143-94大致分為5種，而因本試驗試材不同，破壞模式多出膠合層破壞型。

關鍵字

臺灣櫸；柳杉；補強；膠合樑；結構用材

並列摘要

The study showed Zelkova serrata as the reinforced material for Cryptomeria japonica structural material. First, we measured visual grading method, stress wave method and static bending test to evaluate the laminae. Next, the modulus of elasticity of prediction(Ebp)would be calculated to predict the bending stiffness of reinforced glulams. And the bending properties, shear properties, delaminating properties and longitudinal compressive properties were test as well, in order to analysis and evaluate the properties of the glulams. The bending stiffness data would be calculated by using the modulus of elasticity (MOE)of laminae. The study learned that modulus of elasticity of reality(Ebo)increased when the Ebp increased, and the R2 value was 0.74, which meaned that we could easily predict Ebp based on Ebo. The bending results indicated that using 17 mm Z. serrata reinforced on top side, using 17 mm Z. serrata reinforced both on top and bottom sides, using 23 mm Z. serrata reinforced both on top and bottom sides, had significant differences on reinforcing effect if we only considered about MOR. Among these three samples, sample that using 17 mm Z. serrata reinforced both on top side used the less Z. serrata, could lower the cost. Ebo in this sample were 9.7 GPa, which was 14.5% higher than C. japonica, and MOR were 74.8 MPa, which was 49.8% higher than C. japonica. But when we considered about Ebo and MOR, the bending results indicated that only using 17 mm and 23 mm Z. serrata reinforced both on top and bottom sides have significant differences on reinforcing effect. And compared to these two samples, using 17 mm Z. serrata reinforced both on top and bottom sides could lower the cost. Ebo in this sample were 14.6 GPa, which was 29.2% higher than C. japonica, and MOR were 79.3 MPa, which was 58.8% higher than C. japonica. Furthermore, the smaller samples had similiar result when we considered about Ebo and MOR. Using CNS 11031 standard to evaluate the shear strength, wood failure percentage and delamination rate. The results revealed that some adhesive lines were incomplete. In the research, the reinforced materials only suit for type Ⅱ and type Ⅲ use environment. According to the results of the above-mentioned, it can be found that the gluing property of the test gluing layer is not good, and the reason is related to the glued process and the shrinkage of the material itself. Therefore, in order to improve this phenomenon, we could improved the process or replace other materials of similar dimensional stability. Under longitudinal compression test, the longitudinal compressive properties in each samples did not have significant differences. Using ASTM D143-94 standard to observed the types of failure in longitudinal compression. There were 5 kinds of failure types, including the type that break at the adhesive line.