Evaluation of two anchor plate capacity models for maw systems

Abstract

The present paper examines the accuracy of two models used to predict the capacity of single square plate anchors used in multi-anchor wall (MAW) systems constructed in Japan. Measured anchor plate capacities from full-scale in situ load tests reported in the literature are compared to predicted values using the analytical model recommended in Japan. A modified anchor capacity model is proposed that preserves the general form of the PWRC equation, but introduces correction factors to improve accuracy. The correction factors are empirically-based and are selected by back- fitting measured loads to achieve a resistance bias mean equal to one and a low coefficient of variation (COV) in bias values. The results of a large number of small-scale anchor capacity tests carried out in pullout boxes were used to guide the selection of back-fitted parameters. The correction factors account for the influence of the confining pressure (anchor depth), the plate size and the anchor rod length on the anchor plate capacity. New N c and N q factors are also presented as part of the new model. The bias statistics for the two models comprise an important contribution to future reliability-based calibrations of the ultimate anchor capacity limit state for MAW systems in Japan.