Evaluation of Site Variability Effect on the Geotechnical Data and Its Application

Abstract

Abstract This study focused on evaluating the laboratory measurement errors and site variability due to soil deposition and operator using the following tests: unconsolidated-undrained (UU), Atterberg limits, small direct shear, and California bearing ratio (CBR) tests. Four clay soils with different plasticity indexes (PIs) were tested in the laboratory (PI = 11, 21, 38, 53), in addition to sandy soil and 3 aggregate base materials. Five different operators tested three specimens for each soil type to investigate the effect of operator on testing results. The maximum coefficient of variation (COV) data due to operator were 10.9 %, 7.8 %, 18.2 %, 16.9 %, 19.5 %, and 15.6 % for UU, liquid limit, plastic limit, plasticity index, direct shear, and CBR tests, respectively. The effect of operator and equipment on field testing results was evaluated using 28 cone penetration tests (CPTs) points performed in the same site at different times by two contractors. The spatial variability of CPT data was evaluated for each contactor and both combined. The values of COV for CPT data were different for the two contractors. The average COVs of pile design method and site variability were used to calibrate the Load Resistance Factor Design (LRFD) resistance factor, Φ, which showed high value for lower variability data, and vice versa for high variability data. The operator and equipment variabilities caused Φ to reach 0.24 for CPT data. In addition, the standard penetration test (SPT) and unconfined compressive strength test (qu) were collected from 21 borings and used to evaluate site variability. The value of Φ increased from 0.41, when COV of pile design method was only included, to 0.55, when site variability was included for SPT data. The same scenario was done for qu, in which Φ was increased from 0.34 when including COV of pile design method only to 0.39 when using the average COV of both method and site.