Insight into centrifuge modelling errors in predicting embedment depths of dynamically installed anchors

In centrifuge modelling of dynamic events, a common practice is to preserve the inertial effect. This will lead to a loss of similitude in the strain rate between model and prototype. For centrifuge tests of dynamic installation of anchors, such strain rate dissimilarity may give rise to significant discrepancy in the anchor embedment depth between the model and the prototype, i.e., modelling error. Large-deformation finite element analyses that simulate both the model and the prototype were undertaken in this study to provide insight into this phenomenon. The calculation results revealed that the centrifuge model tends to underpredict the anchor embedment depth of the prototype. A parametric study of the influence of anchor geometry and weight, soil strength and stiffness, strain softening and rate parameters, and centrifuge acceleration level was subsequently conducted. It was found that such underprediction depends heavily upon the strain rate parameter. Based on analyses results, practical advices were given for centrifuge modellers to minimize the modelling error. The results presented and advices given in this study may facilitate the design of centrifuge model tests especially when the anchor embedment depth is an essential detail to be replicated in these tests.