Experimental and Numerical Investigation of Geopolymer Aggregate Concrete Behaviour

Geopolymer Aggregate (GPA) is a novel coarse aggregate synthesized from low calcium fly ash extracted from an Australian power plant with a highly alkaline activator. It is also classified as a lightweight aggregate having a density of 1709 kg/m3. The production of alternative aggregates is an area of study that is contributing to achieving the goal of producing sustainable concrete. However, a thorough understanding of the GPA and the influence of the GPA on the characteristics of concrete is required prior to its application in construction. Due to the lack of long-term experimental data, the mechanical and durability characteristics of GPA concrete beyond 90 days are still unknown. Furthermore, only limited research has been carried out on understanding the variations in mechanical and durability characteristics in relation to the changes in chemical compositions, microstructure, pore structure, and nanomechanical properties. Moreover, the capability to simulate the performance of GPA concretes using a finite element modelling approach is also unexplored. This study aimed at providing an in-depth understanding of the mechanical and durability characteristics of concrete with 100% of geopolymer coarse aggregates and their variations over the long term. The performance was also benchmarked against the characteristics of conventional aggregate concrete. In addition, numerical simulation of GPA concrete was carried out using various finite element modelling approaches, while identifying their applicability to GPA concrete. A comprehensive literature review was conducted to identify the various production techniques of fly ash aggregates and their effect on the characteristics of concrete. The literature review also identified the numerical modelling techniques used for the simulation of conventional aggregate concretes and similar lightweight aggregate concretes. An extensive experimental methodology was designed to implement a wide range of characteristic testing programs for GPA concrete over the long term while identifying the reasons underpinning the variations. In the first phase of the study, a detailed investigation of the mechanical and durability properties of GPA concrete was carried out. The characteristics of GPA concrete were benchmarked against conventional basalt aggregate concrete. The mechanical properties include compressive, flexural, and splitting tensile strengths, and elastic modulus of GPA concrete which ranged from 42.1 and 50.81 MPa, 4.75 and 5.27 MPa, 3.02 and 3.66 MPa, 20 and 20.5 GPa, respectively within the 90-day to 365-day period. The correlations between existing concrete standards and the major mechanical properties of GPA concrete were discussed. Empirical relationships are developed between compressive strength and mechanical properties including flexural strength, splitting tensile strength, and elastic modulus using statistical regression analysis. The suitability of using the existing relationships in Australian standards and American Concrete Institute codes for GPA concrete are critically examined. The long-term creep tests on GPA concrete identified the 1-year creep strain as 747 microstrains while the calculated creep coefficient was 0.97, which is significantly lower than the creep coefficient predicted by AS 3600 and CEB-FIP models. Moreover, the 365-day drying shrinkage in GPA concrete was identified as 570 microstrains, which is also lower than the maximum permissible limit specified by AS3600. The GPA concrete displayed high water absorption, but lower air and water permeability compared to basalt aggregate concrete. The chloride resistance test identified a chloride diffusion coefficient of 7.15 × 10-12 m2/s which was 8.96% greater than that of basalt aggregate concrete when exposed to sodium chloride solution for 90 days. The performance of GPA concrete under sulphate and the acid attack was investigated for up to two years in relation to strength loss, mass loss, visual appearance, length change, ultrasonic pulse velocity, and pH profile and compared with the performance of basalt aggregate concrete and control water-cured specimens. In 5% Na2SO4 and 1% H2SO4 solutions the performance of GPA concrete was identified to be similar to the basalt aggregate concrete for up to two years. However, in more aggressive 5% MgSO4 and 3% H2SO4 solutions, the GPA concrete showed increased strength loss and mass loss. The investigated mechanical and durability characteristics were correlated and explained through the microstructural analysis with scanning electron microscope imaging, energy-dispersive X-ray spectroscopy analysis, and nanoindentation, pore structure analysis with X-ray computed tomography, and the chemical analysis with X-ray Diffraction and Fourier transform infrared spectroscopy at various ages in the study. Finite Element Modelling (FEM) is an effective tool for predicting concrete performance in various applications. Experimental studies of GPA concrete indicate that failure during uniaxial compression occurs through GPAs and ITZs. Therefore, an accurate mesoscale model can conduct parametric studies, potentially reducing the time and cost associated with experimental analysis. Therefore, in the second phase of the study, state-of-art numerical modelling approaches with advanced material modelling were used to simulate the failure mode, localized damage, and uncertainties in uniaxial compressive strength. The simulations were carried out using both the continuum modelling and mesoscale modelling approaches with the aim of identifying a suitable modelling approach to capture the mechanical damage response of GPA concrete. The continuum modelling was conducted with a 3D finite element model with the concrete damage plasticity model of the ABAQUS software. The continuum model overestimated the compressive strength by 7 to 10.5% however simulated the double cone failure mode observed in GPA concrete. A novel Python script was developed to simulate the geopolymer aggregate distribution at a 2D mesoscale sample domain. The script generates geopolymer aggregate and ITZ phases with arbitrarily shaped polygons. A modified piecewise damage model capable of simulating the compressive and tensile behaviour of GPA was developed as a user-defined material model (VUMAT) for the ABAQUS explicit platform. A script was developed using ABAQUS Python Scripting to automate the development of the 2D three-phase mesoscale model to simulate the uniaxial compression and it was calibrated and verified using laboratory experimental data. The peak compressive strength obtained from the simulations yielded an average of 31.61 MPa and a standard deviation of 1.54, which provided an accurate match with the experimental results. The failure pattern depicts the double-cone failure mode and simulates the initiation and propagation of mechanical damage through ITZ, GPA, and cement mortar phases. Overall, the observations of this study demonstrate the potential of using GPA concrete in various structural applications making it a viable and sustainable alternative to conventional aggregate concrete with the added advantage of reducing the environmental impact by utilizing fly ash from coal-fired power generation.