Abstract:
Traditional interpretation and interpolation techniques have been extensively used in conjunction with geotechnical in-situ teststo evaluate subsurface conditions; however, these approaches, when applied at highly variable geomaterial sites, can lead to high uncertaintiesand inaccurate subsurface characterization. This paper presents the application of a proposed framework that combines geotechnical in-situ testdata and a geostatistical modeling approach with visualization techniques to provide an advanced three-dimensional (3D) site characterizationand visualization of an aging hydraulic fill dam. There were 66 cone penetration tests (CPT) performed along the crest, downstream, andupstream sides of the hydraulic fill dam and properties, such as the soil behavior type index (Ic), effective friction angle (φ0), and undrainedshear strength (Su), were determined using traditional CPT data correlations. Univariate statistics performed on the properties revealed highvariability in the hydraulic fill dam configuration. Anisotropic semivariogram models were developed to incorporate the spatial variability andthe directional anisotropy of soil properties in the proposed site characterization approach. A 3D kriging interpolation was then performedbased on the anisotropic semivariograms to generate 3D geotechnical visualization models of theIc,φ0, andSuof the geomaterials encounteredin the dam. Validation studies were performed by comparing the 3D interpolated soil properties with the ground-truth measured values obtainedfrom laboratory tests. The visualization models based on a geostatistical interpolation approach can facilitate the identification of criticalzones within the dam that could be susceptible to geotechnical hazards. This paper highlights a novel approach combining the in-situ data,3D interpolation kriging analyses, and visualization techniques to evaluate the soil configuration within highly heterogeneous sites such ashydraulic fill dams and mine-tailings sites