Evaluating and Incorporating Site Variability in Different Geotechnical Engineering Applications

Site investigation and characterization of subsurface soil conditions are crucial for geotechnical engineering design and analysis. Geotechnical properties vary inherently from one point to another within the same site. This study focused on evaluating the site variability from cone penetration tests (CPT) and soil boring data, and its implementation into the Load and Resistance Factor Design (LRFD) formulation. The total Coefficient of Variation (COVR,total) for each site was estimated from COVR,spatial (Spatial Coefficient of Variation) and COVR,method (Method Coefficient of Variation) which were then used to calibrate the resistance factors (ϕspatial, ϕtotal) for application to LRFD design. In the Bayesian study, the mean bias and standard deviation of bias were obtained from pile load test database from 34 different sites. In this study, we introduced a new term called Confidence Bias Site Parameter (wb) to put more weight on the new test results. We found that the updated posterior is between the state and new data (giving priority to the new data). Traditional interpolation techniques (e.g. kriging) have been proved inadequate for the uncertainty assessment with the soil variables. To solve this problem, we performed Sequential Gaussian Simulation (SGS) and tried to find the prediction probability at LA1 site (which is 75%) and incorporated it into the LRFD formulation. 126 scenarios were considered to investigate the effect of site variability (by changing COV of friction angle (ϕ), cohesion (c) and unit weight, and the vertical and horizontal correlation length) on the factor of safety (FS). Results show that the FS decreased gradually when coefficient variation values for friction angle, cohesion and unit weight increased. We investigated the effect of the variation of soil parameters (c and ϕ) on the resistance factor and the bearing capacity of shallow foundation. When soil boring locations are getting away from foundation location and the COVc and COVϕ increased, the decreased. On the other hand, the reliability indexes decreased when COV values increased. The effect of site variability on the resistance factor of the deep foundation was also investigated. When soil boring locations are getting away from foundation location and with the increment of COVc and COVϕ the resistance factor decreased. However, when the end bearing is included, increased