以PDC資料分析土壤空間變異性

Abstract

迄今爲止諸多關於土壤液化潛勢的研究中，都離不開對土壤強度參數的獲取途徑和對參數料可靠性的評估。工址調查常以標準貫入試驗(SPT)和現場劈管取樣之物性試驗獲得其相關參數，本文旨在利用PDC設備所得的現地資料進行一維空間變異性之分析，其中包括對模擬模型適用性、模擬方法適用性，以及對辨識趨勢函數的可行性與方法的探討。 透過比對PDC資料和既有SPT資料中的土壤分層情況進行模擬N值與深度之波動尺度，進行多種趨勢函數判別法、模擬法、不同的自相關模型之間擬合成果的比較和探討。最終選出貝葉斯分析方法和Whittle-Matérn模型探討波動尺度，在不取樣的情況下對不同土層的波動尺度估算並分類，為PDC鑽探結果之使用提供一種新的加值應用。經研究發現，PDC所得N值與深度的波動尺度大小可分辨出粘土和砂土。

So far,Many studies on soil liquefaction potential are inseparable from acquiring soil strength parameters and evaluating the reliability of parameter materials. The relevant parameters obtain by standard penetration test (SPT) and physical property test of on-site split pipe sampling in the site survey. The applicability of the simulation method and the discussion on the feasibility and method of identifying the trend function. In order to simulate the scale of fluctuation (SOF) in N value and depth, this paper compares the soil stratification in the Piezo Drive Cone (PDC) data and the existing SPT data. Discuss various trend function discrimination methods, simulation methods, and relevant results between different autocorrelation models. Finally, the Bayesian analysis method and Whittle-Matérn model are selected to explore the SOF, estimate and classify the SOF of different soil layers without sampling, and provide a new value-added application for PDC. The research found that the SOF of N value and depth obtained by PDC can distinguish clay and sand.