Stochastic Simulation of Uplift Load-Displacement Behavior of Helical Anchors in Clays

Recent studies have indicated that the uplift load-displacement behavior of helical anchors installed and tested within the same soil deposits and depths has exhibited a significant amount of variability. Appropriate estimates of displacement should consider the uncertainty associated with both the inherent soil variability and the uncertainty associated with the selected load-displacement model. This paper uses a load test database of uplift loading tests of helical anchors within fine-grained plastic soils to estimate and assess the uncertainty associated with inherent soil variability and model error for the purposes of simulating load-displacement behavior. Two load-displacement models are calibrated and their statistical performance quantified using normalized displacement and normalized load. Fitting of the model curves to the empirical data resulted in highly correlated model parameters. Thereafter, bivariate probability distribution functions were generated to allow adequate joint simulation of the load-displacement model parameters. Calibrated copula models were shown to provide superior estimates of load-displacement model parameters, and the resulting load-displacement curves suitably replicate the observed joint displacement uncertainty.