Optimisation of microstructural properties and mechanical performance of asphalt mixtures

Abstract

Excellent gradation with good strength aggregate morphology and quality binder significantly affect the microstructural properties, which affects the mechanical performance of asphalt pavements. Selection of aggregate gradation with optimised complex geometric morphologies and contact point of interaction of aggregate particles to provide good interfacial interactions between aggregate and asphalt binder, which complies with specification requirements, is a lengthy trial and error procedure. Most successful mix designs were obtained based on the experience of the designer. This study’s first aim is to understand the influence of the chemical properties, texture, shape, strength, and, consequently, the microstructural properties and its resultant mechanical behaviour. The second aim is to optimise gradation and aggregate packing with idealised air voids without testing the microstructure in the laboratory. The materials selected in this study were three types of aggregate: granite 1(GD 1), granite 2 (GD 2) and limestone (GD 3), TK fibre and bitumen pen grade 40/60 and 70/80. MATLAB and Particle Flow Code (PFC3D) were used in this study to achieve and optimise the parameters that influence the stability of the aggregate skeleton, including highly detailed aggregate clumps, aggregates clump size range (gradation), and the packing density by using optimum air voids as an indicator. The virtual aggregate skeleton microstructure was built, and a triaxial test in the PFC3D was used. The anisotropic mechanical properties of the virtual aggregate skeleton, such as the normal contact force and void fabric, were measured using a scan line void method established along the angle 𝜃����� using fabric vectors and were obtainable using PFC3D. Additionally, the asphalt mixtures were prepared from three different aggregate sources to evaluate the influence of aggregates’ physical and chemical properties on microstructural properties. The microstructural properties were captured using a Charge-Coupled Device (CCD) digital camera and analysed using Image Process Analysis System (iPAS) software. The effect of the daily temperature variation of the environment on the corresponding temperature gradient through the depth of the asphalt mixtures layers was assessed. Tests including Marshall stability, wheel tracking, three-point bending beam and double tension-compression at two periods were conducted to investigate the relationship between the structural performance of the asphalt mixtures and their microstructural properties. Additionally, the normal contact fabric tensor (F) was developed to evaluate the interlocking of aggregate in asphalt mixtures, and a correlation with the mechanical performance was performed. Also, all the bitumen grades were modified with various percentages of Teak (TK) fibre contents to evaluate rheological properties through rutting, penetration, thermal conductivity, and multiple stress creep recovery tests. Base and teak modified binders were soaked in the sodium carbonate (NaCO3) salt solutions for different conditioning times to simulate rheological performance in semi-arid areas. The Kruskal-Wallis test, a non-parametric test, was used to statistically study the effect of stress level, and TK addition and the mean wise comparisons at a 95% confidence level were determined. Grey relational analysis was selected to evaluate the asphalt mixture's performance through chemical characteristics, strength, shape, microstructural properties, and modified binder. Parameters that influence the microstructure stability, such as aggregate size fraction range, shape percentage, Dominant Aggregate Size Range (DASR), aggregate densities, and Destructive Factor (DF), were optimised for all gradations using PFC3D. The results suggest a linear correlation between the evolution of normal contact fabric and void fabric. A scan line approach successfully identified and quantified the wall and loosening effects by quantifying the voids’ shape and orientation. The results showed that aggregate’ chemical oxides such as sodium oxide (Na2O) + potassium oxide (K2O) + calcium oxide (CaO) + silicon (SiO2) have a major effect on the electrical conductivity of granite samples at high temperatures and pressure. The results indicated that the higher amount of sodium oxide (Na2O) + potassium oxide (K2O) + calcium oxide (CaO) + silicon (SiO2) in the aggregate, the better the aggregate texture and the higher the number of contact points. In addition, high numbers of contact points have a notable impact on the Marshall stability, rutting resistance, dynamic stability, stiffness modulus and complex Poisson’s ratio at high and low temperatures. A good linear correlation between F and Marshall stability, dynamic stability and stiffness modulus was demonstrated. TK fibre modified bitumen provided greater rutting resistance, lower penetration, and lower thermal conductivity than the base bitumen binders. Compared to base bitumen, TK modified bitumen demonstrated reduced shear strain, higher recovery, and higher non-recoverable creep compliance. The TK fibre-modified bitumen binders had better resistance to the action of salt than the base bitumen. However, increased shear strain, reduced recovery, and increased non-recoverable creep compliance were observed for samples soaked in increasing salt concentration. The base bitumen and TK modified bitumen showed excessive softening and negative recovery at stress levels from 0.5k Pa due to increased soaking duration and salt content. The fabric tensor presented in this study is another approach that can evaluate the packing of aggregate in asphalt mixtures. The study concluded that aggregate contact properties directly influence the Marshall, rutting, and bending stiffness modulus for both low- and high-temperature performance of asphalt mixtures. The results for F and the number of aggregate contact points were similar. Therefore, F can also be used to evaluate the internal structural index. This investigation established that adding TK fibre to bitumen enhanced its toughness, which could improve asphalt fatigue performance, with the optimum level of TK fibres being 2% by mass of bitumen.