動態時間域之鋪面反算分析與研究

Dynamic Time Domain Backcalculation Analysis of Pavement

落體測位儀試驗已廣泛的用於現有鋪面結構強度與績效之評估。目前反算分析方法大都以靜態假設為主，此方法是以試驗所得各感應器最大位移資料反算得鋪面材料性質。然而落體測位儀試驗為動態試驗且亦有完整之時間歷時資料，若能使用動態理論進行分析將可更合理的模擬鋪面之實際反應。本研究首先將動態有限元素程式與最佳化理論結合發展時間域動態反算程式，並使用人造動態歷時資料進行程式之驗證。時間域動態反算程式之優點為可避免試驗資料錯誤之影響及能找出感應器與地面之耦合效應。接著再將所發展反算程式分析美國公路總署之長期鋪面績效(Long-Term Pavement Performance, LTPP)資料庫中六個試驗路段之落體測位儀試驗資料，並將預測結果與試驗值作比較。最後亦將反算所得材料參數預測其它不同落錘荷重下之動態反應，結果說明本研究反算所得材料參數可合理的預測鋪面在其它落錘荷重作用之反應。此外，亦將動態反算結果與靜態反算作比較。

Falling Weight Deflectometer has been used to evaluate the structural capacity of in situ pavement. Most backcalculation analayses are based on the static assumption, and the pavement layer moduli are determined from the peak deflection of FWD test data. However, the FWD test imparts a dynamic load, and the time history data are recorded. The dynamic backcalculation can use these data and predict the FWD response. A dynamic finite element program is incorporated with an optimization approach to develop a time domain backcalculation program, and the developed program is verified by simulation. The advantage of a time domain analysis is that it can use the time history data directly and is less affected by the error data and the coupling effect between the sensor and asphalt concrete. Finally, the backcalculation program is used to analyze six sections of FWD data which are obtained from the FHWA Long-Term Pavement Performance database in the U.S., and the backcalculated properties are used to predict the dynamic response of the FWD test in which the deflection time history data is not used for backcalculation analysis. The results show that the backcalculated properties can predict the dynamic response of a FWD test reasonably. Finally, the comparison of the dynamic and static backcalculation is studied.