Design methodology for site-specific resistance factors based on foundation location and size

摘要

In reliability based design for deep foundations, design provisions specify resistance factors for a multitude of design methods, foundations types, and methods of field verification (e.g., load testing). The prescribed resistance factors are based on specified target reliabilities and are generally calibrated using databases of measured and predicted resistances. A limitation of this approach is that it does not account for particular degrees of variability of material properties that may be encountered across different sites (i.e., homogeneous versus heterogeneous). The present work presents a methodology that upholds the framework of reliability based design while accounting for site-specific spatial variability and maintaining consistency with current practice in deep foundation design. Presented herein are geostatistical tools that can be used to characterize site heterogeneity and quantify uncertainty of foundation resistance through stochastic simulation. Both uncertainty reduction due to conditioning to nearby samples and those uncertainties associated with incomplete site characterization are incorporated into the procedures. An additional uncertainty, which is not related to spatial variability of site properties is termed "method error" (e.g., measurement, transformation, construction errors) and is accounted for using a database of load test results. Implementation examples are shown for multiple cases of driven pile and drilled shaft foundations.