基於CPT資料庫統計自協方差參數

Development of new statistics for auto-covariance parameters based on CPT database

許曉鈞(Xiao-Jun Xu)

指導教授 ： 卿建業

國立臺灣大學/工學院/土木工程學研究所/碩士(2022年)

https://doi.org/10.6342/NTU202201517

摘要

在近年的研究中發現，Ching et al.(2019)指出樣本路徑平滑度和Chang et al.(2021)孔洞效應也被視為重要的參數，樣本路徑平滑度與孔洞效應對失效概率影響顯著。由於這兩個重要參數的發現，前人CPT資料庫中分析得到的自協方差參數必然存在誤差。在實踐中，設計工程師經常根據文獻中發現的類似場址的SOF統計資料採用合理的SOF。這些統計資料很有價值，因為對於現場調查資料不足的設計案例，這些統計資料可用於根據其土壤類型獲得一階SOF估計。所以用最可靠的模型與方法分析與統計出最新的CPT資料庫自協方差參數十分重要的，這些新的統計資料對於場地調查資料不足的未來設計可能很有價值：它們可用於獲得基於土壤行為類型的自協方差參數的一階估計。
本次研究中採用Whittle-Matérn模型以及Cosine Whittle-Matérn (CosWM)，估算方法採用的是Maximum likelihood method(ML)(最大似然法)。並通過統一標準對資料庫中的錐尖阻抗(qt)資料進行分析，以確定其自協方差參數，包括變異係數、關聯性長度SOF、樣本路徑平滑度和孔洞效應。
本文編制了一個圓錐貫入試驗(CPT)資料庫，命名為Global-CPT/1138(59)，基於識別的自協方差參數的統計得出新結論。SBT等於3 (clay)和6 (sand)的變異係數(COV)小於SBT等於4 (silt mixtures)和5 (sand mixtures) 的變異係數(COV)。
對於SBT等於3 (clay)、4 (silt mixtures)和5 (sand mixtures)，大多數情況下孔洞效應不顯著(模型選擇WM)。對於SBT等於 6 (sand)，大多數情況具有顯著的孔洞效應(模型選擇CosWM)。Clay-like soils(SBT等於3或4)的SOF範圍相對較窄(大多為0.1m-0.5 m)，而砂狀土壤(SBT等於5或6)的SOF範圍相對較寬(大多為0.06 m-1.3 m)。所有土壤的樣本路徑平滑度範圍都很窄(幾乎為0.1-1.3)。

關鍵字

關聯性長度 ； 樣本路徑平滑度 ； 孔洞效應

並列摘要

In recent studies, Ching et al. (2019) pointed out that the sample path smoothness and the hole effect of Chang et al. (2021) are also regarded as important parameters, and the sample path smoothness and the hole effect have an impact on the failure probability. Not significant. Due to the discovery of these two important parameters, there must be errors in the auto-covariance parameters analyzed in the previous CPT database. In practice, design engineers often employ a reasonable SOF based on SOF statistics found in the literature for similar sites. These statistics are valuable because they can be used to obtain first-order SOF estimates based on their soil type for design cases with insufficient field survey data. Therefore, it is important to analyze and statistic the latest CPT database auto-covariance parameters using the most reliable models and methods. These new statistics may be valuable for future designs with insufficient site survey data: they can be used to obtain soil behavior based a first-order estimate of the auto-covariance parameter of type.
In this study, the Whittle-Matérn model and the Cosine Whittle-Matérn (CosWM) model were used. The estimation method was the Maximum likelihood method (ML) (maximum likelihood method).
In this study, a database of cone penetration tests (CPT), named Global-CPT/1138(59), was compiled to draw new conclusions based on the statistics of the identified autocovariance parameters. The coefficient of variation (COV) for SBT equal to 3 (clay) and 6 (sand) was less than the coefficient of variation (COV) for SBT equal to 4 (silt mixtures) and 5 (sand mixtures). For SBT equal to 3 (clay), 4 (silt mixtures) and 5 (sand mixtures), the hole effect is not significant in most cases (model choice WM). For SBT equal to 6 (sand), most cases have significant hole effects (model choice CosWM). Clay-like soils (SBT equal to 3 or 4) have a relatively narrow SOF range (mostly 0.1m-0.5 m), while sandy soils (SBT equal to 5 or 6) have a relatively wide SOF range (mostly 0.06 m-1.3 m). The sample path smoothness range for all soils is narrow (almost 0.1-1.3).

並列關鍵字

scale of fluctuation ； sample path smoothness ； hole effects ； Whittle-Matérn ； Cosine Whittle-Matérn ； maximum likelihood method

參考文獻

Abramowitz, M. and Stegun, I. (1970). Handbook of Mathematical Functions. Dover, New York.

Akkaya, A. D., and Vanmarcke, E. H. (2003). Estimation of spatial correlation of soil parameters based on data from the Texas A M University NGES. Probabilistic site characterization at the National Geotechnical Experimentation Sites (Eds G. A. Fenton and E. H. Vanmarcke), 29-40.

Benjamin, J.R. and Cornell, C.A. (1970). Probability, Statistics, and Decision for Civil Engineers. McGraw-Hill, New York.

Bong, T. and Stuedlein, A.W. (2017). Spatial Variability of CPT Parameters and Silty Fines in Liquefiable Beach Sands. Journal of Geotechnical and Geoenvironmental Engineering, 143(12), 04017093.

Bong, T. and Stuedlein, A.W. (2017). Effect of cone penetration conditioning on random field model parameters and impact of spatial variability on liquefaction-induced differential settlements. Journal of Geotechnical and Geoenvironmental Engineering, 144(5), 04018018.