加勁擋土結構（reinforced soil structures或稱mechanically stabilized earth walls）於1980年代後期引進台灣，由於高分子聚合物的快速研發，各式加勁材料及加勁擋土結構系統也迅速地蓬勃發展。其優點在於可大量減少施工經費、縮短工期、減少不均勻沈陷、容許較大變形、耐震性能較佳及可植生美化景觀等。根據不同破壞面形狀與土壓力分佈的假設，目前已發展出許多加勁擋土牆的設計方法，但是針對相同的加勁擋土牆，不同方法的設計結果差異甚大，因此如何評估牆體臨界破壞的狀態是相當重要的議題。此外，台灣的山坡陡峻，加勁邊坡之坡面傾斜度較高，高度亦常達20~40公尺，超過一般國際慣用高度，因此設計時更需小心謹慎。現有規範針對雙階式加勁擋土牆（superimposed geosynthetic reinforced earth wall）的設計規定，多未以全尺寸試驗或物理模型試驗的方式加以驗證，因此，不同退階距離（offset distance）和加勁材料長度對牆體穩定性與破壞面位置的影響是需要加以研究的。 本研究以中央大學地工離心機進行離心模型試驗，探討加勁擋土牆的斷裂破壞行為，並由試驗結果進行內部穩定分析。本研究之目的有三個：（1）釐清加勁材料強度在離心模型試驗中的尺度因子；（2）對加勁擋土牆的內部斷裂破壞提出新的評估方法，使設計結果能更接近破壞臨界的狀態；（3）檢核美國聯邦高速公路管理局（FHWA）對雙階式加勁擋土結構的設計方法。 本研究進行18組單階式加勁擋土牆與29組雙階式加勁擋土牆之離心模型試驗，由試驗結果可以得知：（1）本研究所提出之修正側向土壓力設計法（modified lateral earth pressure design method），可根據加勁擋土牆的幾何條件與土壤性質，準確的評估加勁擋土結構的加勁材料強度與加勁間距，使設計結果接近臨界破壞的狀態。（2）本研究所提出之加勁擋土結構牆體穩定評估法（RESS assessment），可根據所選用的加勁材料強度、加勁間距與土壤條件，準確的評估加勁結構的臨界高度，使設計結果接近臨界破壞的狀態。（3）美國聯邦高速公路管理局對雙階式加勁擋土牆的設計規定需進行修正：a. 當退階距離小於 (H1+H2)/6.8時，雙階式加勁擋土牆可視為單階式加勁擋土牆設計；b. 當退階距離介於(H1+H2)/6.8與H2cot(45+φ/2)之間時，雙階式加勁擋土牆可視為複合式加勁擋土牆設計；c. 當退階距離介於H2cot(45+φ/2) 與H2tan(90-φ)之間時，上階可視為單階加勁擋土牆設計，下接則視為複合式加勁擋土牆設計；d. 當退階距離大於H2tan(90-φ)時，雙階式加勁擋土牆可視為獨立兩個單階加勁擋土牆分別設計。（4）以顆粒性土壤為背填材料之加勁擋土牆破壞時，其加勁材料強度在離心模型試驗中的尺度因子為1/N。（5）不論是單階式加勁擋土牆還是雙階式加勁擋土牆，傳統平面破壞面（破壞面夾角(β’+φ)/2）即可完整包絡破壞區域。設計時使用傳統平面破壞面，可簡化美國聯邦高速公路管理局之加勁擋土結構的設計與施工規範中對雙階式加勁擋土牆破壞面位置的決定方式。（6）試驗結果的回饋分析顯示，針對8層至16層的加勁擋土牆，其內部側向土壓力係數會隨著加勁層數的增加而減小；當加勁層數大於16層時，土壓力係數則趨於定值。 The purposes of this study are threefold: (1) clarify the scaling factor of the reinforcement strength; (2) propose the assessment methods for evaluating internal stability of the geosynthetic reinforced earth structures to close to the verge of failure; (3) examine the empirical design rules for the design of superimposed geosynthetic reinforced earth wall (SGREW) in FHWA guidelines. To meet these purposes, a total of 47 models of geosynthetic reinforced earth structures were tested using geotechnical centrifuge in this study. From this study, two convenient assessment methods were proposed enabling the accurate evaluation of the internal stability of geosynthetic reinforced earth structures on the verge of failure. From the observation and the analysis of the modeling test results, conclusions are (1) the proposed modified lateral earth pressure design method is for evaluating the required reinforcement spacing or reinforcement strength for geosynthetic reinforced earth structures by the use of the factor of Rf which increases with increasing wall inclination and reinforcement layers; (2) the proposed RESS assessment is for estimating the critical wall height for a geosynthetic reinforced earth structures by the use of factor of Am which increases with increasing wall inclination and reinforcement layers; (3) for a SGREW, if the offset distance is smaller than (H1+H2)/6.8 instead of (H1+H2)/20 stipulated in FHWA guidelines, it should be designed as a single wall; if the offset distance is between (H1+H2)/6.8 and H2cot(45+φ/2), it should be designed as a composite wall; if the offset distance is between H2cot(45+φ/2) and H2tan(90-φ), it should be designed as two single walls with different reinforcement lengths; if the offset distance is greater than H2tan(90-φ), it should be considered as two single independent walls; (4) in centrifuge modeling, the scaling factor of reinforcement strength for geosynthetic reinforced earth structures at failure is 1/N; (5) the failure plane with angle of (β’+φ)/2 can envelop the whole fractures of reinforcement and fit well to the failure wedge, where β’ is the equivalent wall inclination and φ the soil friction angle. Thus, the failure plane angle of a SGREW can be determined using (β’+φ)/2 instead of using complicated factors in FHWA guidelines; (6) for a given geosynthetic reinforced earth structure, if the reinforcement layers are between 8 and 16, the modified lateral earth pressure coefficient Ka’ decreases with decreasing wall inclination and increasing reinforcement layers; however, if the reinforcement layers are greater than 16, the modified lateral earth pressure coefficient Ka’ remains constant.
