Use of Indirect Tensile Test for Developing a Model for Predicting the Fatigue Life of Asphalt Concrete

Fatigue cracking is one of the major forms of distress that directly influences the service life and ride quality of pavements. The fatigue resistance of asphalt mix is defined as the ability of asphalt mix to withstand repeated bending without fracture. A traditional approach to dealing with cracking in asphalt pavements is based on the assumption that cracks initiate at the bottom of the asphalt layer due to tensile stresses developed from the flexure of the layer and propagate to the pavement surface under repeated load applications(bottom-up cracking). However, recent field studies suggest that fatigue cracks may also initiate at the pavement surface and propagate downward under traffic (top-down cracking).Several methods have been developed for the fatigue testing of asphalt mixes that repeated flexural beam, indirect tensile or direct tensile tests are usually used. Since determination of this characteristic is time consuming and it is difficult to measure, some predicting models have been developed. The results of these models are usually differ from each other because of material types, test methods and tests conditions. In general, there are two approaches to analyze and design against fatigue failure: conventional (traditional) approach, which bases on the analysis on the nominal (average) stresses in the region of the component being analyzed; and fracture mechanics approach, which specifically treats growing cracks using the method of fracture mechanics. In recent years other approaches such as viscoelasticity and continuum damage, has been considered. In this paper, based on traditional approach and by using indirect tensile test (for measuring the fatigue life and resilient modulus) a simplified model is developed for estimation of fatigue life of common asphalt mixes which are ordinary used in Iran. The loading pattern used in the indirect tensile fatigue test was a haversine load. The loading time was 0.1-second, and the rest period was 0.9-secend. The amplitude of the load for a specific tensile stress was kept constant during the test. The asphalt mix variables were bitumen content, bitumen type, air void content and aggregate gradation. For resilient modulus and fatigue tests, the Universal Testing Machine (UTM) was applied. The statistical analysis for each data set (resilient modulus and fatigue response) included the following sequences:


Test for correlation among the independent variables.

Analysis of variance (ANOVA) of full models (all main factors and two-factor interactions) to determine the sensitivity of resilient modulus and fatigue life to mix variables.

General Linear Modeling (GLM) to develop models for resilient modulus and fatigue life.

Summarizes of the effects of the experimental variables included in the experiment on resilient modulus and fatigue life based on the results of GLM.
The experimental design used in this study was a fractional factorial which permitted the estimation of the main effects of experimental factors and all the two-factor interactions. On the basis of the statistical analysis the following equation was obtained:R2=0.800 Where:Stress = tensile stress ( KPa)Mr = resilient modulusVFB = voids filled with bitumenFigure 1 shows the fatigue life for specimens resulting from proposed model and indirect tensile fatigue test. As shown in this figure, the predicted values and observed values(results of fatigue tests) are near to each other.