AASHTO自1986年提出T274回彈模數試驗規範以來，至今歷經多次的修正與檢討及至目前最新的T292(1998)規範，其因於回彈模數試驗結果的變異性偏高且無有效方法能加以評估。本研究利用人造試體為校正基準，分別建立人造試體及三軸試驗系統量測不確定度分析模式，並以ISO建議之程序評估系統不確定度。使用人造試體校正三軸試驗系統，能有效界定回彈模數試驗系統所產生的不確定度，並可以此作為判斷試驗結果合理與否的依據。量測不確定度分析結果顯示，因LVDT之量測範圍極小，故對試驗結果可能引致之變異較大。而LVDT對回彈模數三軸試驗系統不確定度影響最大，荷重元次之。 以重覆加載方式對凝聚性土壤進行永久變形試驗，可依應變增加率界定其視搖墜應變產生時的加載次數及永久變形量。而以視搖墜應變量為5%時所建立的路基土壤永久變形破壞應變上限，能評估路基土壤於特定應力水準及含水量組合條件下產生永久變形破壞的可能。同時，建立路基土壤永久變形、回彈應變與視搖墜應變的迴歸模式，可對土壤受反覆載重下的變形及回彈特性提供迅速而有效的預測方式。 Since 1986, AASHTO standardized the test method for subgrade soil resilient modulus in specifications. Still the skills for the operation of the equipment was lacked, and variation and accuracy in test results remain difficult to control. In this paper, synthetic specimens were used to calibrate the triaxial testing system. Evaluation on the uncertainty analysis showed that the precision and accuracy of linear variable displacement transformer has the greatest effect on resilient modulus test results; while that of load has less effect. Results derived from uncertainty analysis on the test system can be used to bound the upper and lower limit of the test outcome, such that criteria for correction or rejection of resilient modulus test results can be established. Permanent deformation of cohesive soil with repeated loading acted can be boundary with shakedown deformation and the loading cycle. Set 5% as failure strain upper limit can evaluate the possibility of subgrade soil under specific stress level and water content situation. And also, regression models of resilient and permanent strain shakedown strain was established and provided a effect way to prediction.
