Numerical Modelling of Cyclic Pile Capacity in Clay

Abstract A numerical procedure for evaluation of the response of a single pile in clay under static and/or cyclic axial loading was developed. Under combined static and cyclic loading, two separate but compatible models representing respectively the static and cyclic responses are established. The procedure is fast and efficient as it attempts (o model an equivalent pile response for the entire design storm, rather than follow the pile response through every load cycle. Introduction Significant efforts were made during the past two decades to improve the methods for evaluating the axial capacity of offshore piles (e.g. Refs. 1-5). As a result of these efforts, the 17th edition of API RP2A (Ref. 6) design guidelines completely revised the recommended procedure for evaluation of axial pile capacity in clay. Furthermore, in the commentary to API RF2A (17th edition or later), the necessity of separately evaluating the effects of pile length flexibility and cyclic loading on the pile capacity are pointed out. During the period 1980-1990, the Norwegian Geotechnical Institute (NGI) performed several comprehensive pile testing programs in overconsolidated (Refs. 7-9) and soft, normally consolidated clay deposits (Refs. 10-12). The tests covered a wide range of static and cyclic loading combinations and provided the basis for the numerical model described in this paper. A number of factors should be considered in modelling of cyclic axial loading of piles, The most important of these factors are:type of cyclic loading (one-way vs. two-way, loadcontrolled vs. displacement-controlled) and number of load cycles,soil properties and variation of soil strength andstiffness with depth and in the lateral direction,pile length and flexibility,initial stress distribution along the pile under static loads prior to cyclic loading, and 5) compatibility in terms of cyclic and average displacements and stresses. The numerical model described below for analysis of static and/or cyclic axial loading of offshore piles accounts for aH these factors. This paper includes the results from recent verification and case studies by NGI and Det Norske Veritas. In these studies, the prdlcted and measured behaviors of instrumented large diameter and model piles are compared. The paper also includes a comparison between the developed numerical model and the RATZ program developed by Randolph (Ref. 13). Soil-Pile Interaction The soil-pile interaction model is based on the so-called concept (Fig. 1): the mobilized shear force t, around a given pile segment is only assumed to be a function of the vertical displacement z, of the mid-point of the segment. The t-z springs in the model are non-linear. They are generated by numerical integration in the radial direction of the mobilized shear strains in each soil layer (Fig. 2a).