Year

2008

Degree Name

Doctor of Philosophy

Department

School of Civil, Mining & Environmental Engineering

Abstract

The road pavement is continuously subjected to the loads transferred from moving vehicles during its design life and stresses and strains (responses) are induced in all pavement layers. The responses occur in the subgrade, which is the in-situ soil stratum with a low modulus, would gradually accumulate with the increase in volume of traffic although those responses are low in their magnitudes. It is reasonable to assume that the responses which occur in other pavement layers constructed according to the specifications outlined by the road authorities are recoverable due to their high moduli. Therefore, the performance of a road pavement significantly depends on the behaviour of its subgrade under the action of moving traffic. Moreover, a study on this behaviour of nonlinear subgrade would increase the confidence of pavement engineers in the design of flexible pavements. The in-depth literature review has revealed that the nonlinear behaviour of subgrades under the action of moving traffic has not been incorporated into the design charts that are currently being used. Furthermore, the multi-layer elastic theory has been used for developing those design charts, although, the sophisticated and fairly expensive finite element computer codes are able to produce accurate results. Therefore, the intention of this study is to investigate the nonlinear behaviour of subgrades under cyclic loading and to develop a design chart for flexible pavements incorporating that behaviour of subgrades employing the finite element theory in all the analyses. The granular pavements are the most widely used type of flexible pavements in Australia. Since subgrade materials with CBR values 5 and 7 are commonly found in road corridors of Australia, two granular base pavements tested under the cyclic loading by means of Accelerated Loading Facility (ALF) were selected from Callington, South Australia (with a subgrade of CBR of 5) and Beerburrum, Queensland (with a subgrade of CBR of 7). An appropriate pavement block is selected from the tested area and modelled as a finite element model for each pavement. The applied cyclic ALF load was modelled as a stress pulse with a cycle time and amplitude. The pavement materials, except subgrades, in both pavements were modelled as linear elastic materials with their moduli and Poisson’s ratios. Two curves of stress versus strain were developed using available literature and used to model the nonlinear subgrades in both pavements. Since an analysis with millions of cycles of load applications would require enormous amount of computer time, the analysis was carried out in stages. When an analysis is carried out in stages, the stresses and strains in nonlinear materials are to be submitted at the beginning of each stage as initial conditions. Two curves of permanent strain versus number of load applications were developed using available literature for both subgrade materials. With the aid of those curves together with the developed stress versus strain curves, the stresses and strains in nonlinear subgrade materials at the beginning of each stage were determined. The modelled pavement with the modelled linear and nonlinear materials together with initial conditions was analysed when subjected to the modelled cyclic load in each stage. The maximum deformations computed on the surface of the pavement were compared with their corresponding field measured permanent deformations. The agreement between the computed deformations and the field measured deformations during the ALF testing has revealed that the curves of stress versus strain as well as the curves of permanent strain versus number of load applications developed for two subgrade materials for studying their nonlinear behaviour under cyclic loading produce satisfactory results. Therefore, those curves were employed in estimating the deformations of granular base pavements with subgrade materials with CBR values of 5 and 7. Two sets of granular base pavements consist of identical thin asphalt surfaces and granular bases were considered for the analysis. The CBR values of subgrade materials in these two sets were considered as 5 and 7 respectively. The base thickness of these pavements was changed from 350mm to 750mm with 100mm increments for both sets. The analysis was carried out in stages for each pavement using the general purpose finite element computer code ABAQUS/Standard. The maximum deformation computed at a point on the surface at the end of each stage was plotted against the corresponding number of load applications for each pavement. Those curves were used to estimate the number of load applications that would produce a deformation of 25mm on each pavement. The estimated number of load applications was assumed as the allowable number of Equivalent Standard Axles (ESA’s) applied on each pavement during its design life. The base thickness was plotted against the allowable number of ESA’s for each pavement and presented as the design curves for CBR 5 and CBR 7 in the design chart developed in this study. In conclusion, the agreement between the computed deformations and the field measured deformations during the ALF testing has demonstrated that the curves developed for studying the nonlinear behaviour of subgrade materials under cyclic loading were able to estimate the deformations in real pavement structures quite accurately.

Share

COinS
 

Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.