利用聯合常態分佈模擬CPTU土壤參數

Abstract

現地的調查對於土木工程而言，Piezocone penetration tests(CPTU)在近幾十年來對於收集地下土壤參數來說非常受歡迎；利用Piezocone penetration tests (CPTU)收集針對地下垂直土壤剖面連續性的參數資訊。本研究將探討利用聯合常態分佈模擬CPTU試驗的錐尖阻抗、CPTU試驗錐面後部量測得知超額孔隙水壓力、不排水剪力強度、以及在輕度過壓密黏土中的過壓密比的可能性。 因此首先本研究收集上述在輕度過壓力黏土中的CPTU土壤參數，藉此建立起一個分析的土壤參數為聯合常態分佈的數據庫。此聯合常態分佈可以被使用於模擬分析的土壤參數之間彼此的相關性，以及利用在Bayesian分析中推導出分析中各參數間之相關方程式。 之後將會在於另一個數據庫中收集在輕度過壓密比黏土中上述分析的土壤參數所組成的另一個獨立數據庫，以及利用文獻中前人學著提出的經驗公式，與利用此聯合常態分布模擬分析中的土壤參數之間的相關性，與利用在Bayesian分析中推導出各參數間相關的方程式進一步進行驗證。

This study explores the possibility of modeling CPTU cone tip resistance, excessive pore pressure behind the cone, undrained shear strength, and overconsolidation ratio of lightly overconsolidated clays as a multivariate normal distribution. This is the second attempt made by the authors for modeling parameters of clays as multivariate normal distributions. This study compiles a large database of lightly overconsolidated clays to construct the multivariate normal distribution among the aforementioned soil parameters. This multivariate normal distribution is then used to simulate the correlations between soil parameters of interest and to derive useful equations for Bayesian inference. This constructed multivariate normal distribution and equations are further validated by another independent database of lightly overconsolidated clays as well as by empirical equations proposed in literature.