A METHOD TO PREDICT PILE RESPONSE UNDER VEHICULAR LOADING FOR INTEGRAL ABUTMENT BRIDGES
MetadataShow full item record
The objective of this work is to develop a design method for Integral Abutment Bridges (IABs), such that more accurate predictions of pile response to vehicular loading can be made. This paper presents the results of field tests performed on piles which support an IAB. These results led to a recommended IAB design procedure which includes an in-situ lateral pile load test, the development of P-Y curves, and the establishment of non-linear soil spring constants for modelling IABs. A brief look at the soil-structure interaction (SSI) and mechanics of IABs, as well as a literature review of the evolution of the IAB design process, are included. Early works including the Winkler and elastic continuum theories lead to the establishment of solutions, which provide displacement and stress values used for design purposes. However, the advent of the home computer made Finite Element Modelling (FEM) a more viable and accurate approach to solving “beam on elastic subgrade” problems. A hybrid approach, combining FEM and Winkler theories (FEM-Winkler), has since become standard practice for designers of IABs throughout Canada. An example of such is presented herein. The coefficient of horizontal subgrade reaction, kh, from which are derived the stiffnesses of discrete soil springs used in the hybrid FEM-Winkler design process, is a variable that is difficult to estimate. Therefore, some structural and geotechnical engineers recommend conducting tests to better assess the actual lateral soil spring stiffnesses. Many of these tests, including those described in the American Society for Testing and Materials (ASTM) Standard D3966, are designed to be conducted on expensive mock-up pile assemblies with similar characteristics to those planned for the production piles. An alternate and more economical approach, consisting of an in-situ lateral pile test conducted on a production pile, was conducted. A live load test to measure bridge deck deflection and pile strain was also performed, and both tests are presented herein. Based on the results of the in-situ lateral pile test, structural models were created with the structural software STAAD.Pro, and assumed values of linear lateral soil spring stiffnesses were used to match the pile head deflection and strain measured in the field. Initial estimates of spring stiffness were based on values of the variation of modulus of horizontal subgrade reaction (nh) that were recommended in the Canadian Highway Bridge Design Code (CHBDC) CAN CSA S6-14, and the Canadian Foundation Engineering Manual (CFEM,1992) for each of three (3) soil strata. These soil spring stiffnesses were found to be accurate at high soil strain, but much too flexible at low strain, which indicated soil non-linearity. Therefore, non-linear reaction-deflection (P-Y) curves were developed from the linear soil spring models, by noting the soil reaction at each discrete soil spring location, based on modelled soil/pile deflection. Non-linear soil springs were defined from the tangential stiffness of the P-Y curves, and these were used to refine a three-dimensional FEM-Winkler bridge model. A comparison of the model incorporating non-linear soil springs with the model incorporating linear soil springs demonstrated that each had a similar response to vehicular loading. Further to this, the predicted pile stresses and deck deflections from both bridge models agreed with values measured during the vehicular live-load bridge test.