| As part of an overall research project dealing with the mechanistic evaluation of Georgia Loaded Wheel Tester (GLWT) as used in assessing the rutting performance of hot asphalt concrete mixtures (HACM), this dissertation is focused on three specific objectives: (i) To provide a detailed study of the predictive capabilities of a new constitutive model (by Saleeb), together with its material parameters estimations, for comprehensive anisotropic viscoelastoplasticity damage representation of different HACM, (ii) To implement a number of robust algorithms for the long-term inelastic analysis (hundreds of thousands of load cycles); and (iii) To perform a detailed simulation of complex interaction in wheel-Hose-HACM specimen.; With regard to item (i) above, the internal-variable formalism, within the general frame work of thermodynamics of interacting dissipative mechanisms, has been utilized in the derivation, and a very robust, unconditionally-stable, integration algorithm was implemented to handle the stress update. On the modeling side, the extreme versatility of this representation has been demonstrated by comparisons to experimental results in a rather elaborate test matrix involving both material and structural (GLWT) testing for different mixes under wide variety of level-and-conditions (temperatures, load rate, stress level-and-triaxiality, varying orientations of specimen-to-compaction axes, cyclic ratchetting and mobility, peak-and post-peak responses, etc).; Considering item (ii) above the key to the development of long-term simulation tool for GLWT software is the ability to rapidly perform the calculations-per-cycle and simultaneously minimize the input/output overhead involved in the use of commercial codes such as ABAQUS. To this end, several routines and script files have been developed as part of the GLWT code for optimization (i.e., to handle the three major tasks involved in primal analysis, sensitivity analysis, and nonlinear optimization), thus enabling the simulations of more than hundred thousand cycles.; Finally, in connection with item (iii), the final outcome has been the significant gain in mechanistic insight of GLWT. In particular, this has led to development of a very efficient, but quite simplified, computer modeling technique for the GLWT configuration (termed sectional analysis) to replace the straight forward, but complicated (and time-consuming) counterpart employed by others; i.e., the three-dimensional/contact model for wheel-hose-asphalt specimen/moving wheel-load analysis. |
