Cement-based materials are the most widely used man-made materials in the world. The durability of cement-based materials has been a major concern due to the premature failure and serviceability issues of many reinforced concrete structures. Durability of cement-based materials is to a large content dependent on their resistance to the ingress of aggressive species, such as chloride, sulfates and carbon dioxide, which is governed by different mechanisms including permeation, absorption, diffusion, or their combinations. Therefore, the primary factors governing the durability of reinforced concrete structures are the transport properties in cement-based materials, e.g. permeability and diffusivity, which depend on the evolution of their underlying microstructures over a wide range of length scales, concentration of aggressive species and environmental conditions, such as temperature, humidity and their variations. The main objective of this thesis is to develop a multiscale modelling scheme to estimate the transport properties in cement-based materials capturing the structural information at each scale and taking into account the influences of w/c ratio, time, chloride binding, degree of water saturation, chloride concentration, interfacial transition zone (ITZ), and aggregate content and shape. To achieve this objective, based on the homogeneous multiscale method, the hybrid lattice Boltzmann-finite element method is proposed and developed in this thesis, which combines the advantages of both lattice Boltzmann method and finite element method. Lattice Boltzmann method is used as micro-scale solver for predicting the transport properties in cement paste and ITZ. Finite element method is selected as meso-scale solver to estimate the transport properties in mortar and concrete. The upscaling between the micro-scale and meso-scale simulations is accomplished by using the volume averaging technique. The representative elementary volume (REV) of cement-based materials at a lower scale is determined with a numerical-statistical approach. Transport properties in the REV of cement-based materials at a lower scale are considered as input to predict the transport properties in cement-based materials at a higher scale. The developed multiscale lattice Boltzmann-finite element modelling framework enables to acquire a meso-scale solution, i.e. transport properties, while still capturing the micro-scale information. The methodology for multiscale modelling of transport properties in cement-based materials is presented in Part II of this thesis, made up of Chapters 3, 4 and 5. In Chapter 3, an overview of the implementation of multiscale modelling scheme is given. Chapter 4 describes the details of the micro-scale solver: lattice Boltzmann method, consisting of the PermLBS (the abbreviation of “Permeability Lattice Boltzmann Simulation”) module, SCMPLBS (the abbreviation of “modified Shan-Chen Multi-Phase Lattice Boltzmann Simulation”) module and DiffLBS (the abbreviation of “Diffusivity Lattice Boltzmann Simulation”) module. A series of benchmark tests are performed to validate each module. Chapter 5 presents the procedure to obtain the 3D structures of cement paste, ITZ, mortar and concrete. At micro-scale, the microstructures of cement paste are obtained through the cement hydration and microstructure formation model, HYMOSTRUC3D model, as well as the X-ray computed microtomography (?CT) scan. Cement paste is regarded as a heterogeneous material composed of capillary pores, hydration product and unhydrated cement grain. At meso-scale, both mortar and concrete are modelled as three-phase composite materials consisting of matrix, aggregate and ITZ. For the mesostructure of mortar, the aggregate is sand, and the matrix is bulk paste, while for the mesostructure of concrete, the aggregate is coarse aggregate (e.g. crushed stones, river gravels), and the matrix is mortar. The microstructure of ITZ is extracted from the “ribbon paste”. The microstructural evolution of the ITZ is simulated with the HYMOSTRUC3D model taking into account the wall effect. The results of multiscale modelling of transport properties in cement-based materials are shown in Part III of this thesis consisting of Chapters 6 and 7. Chapter 6 presents the micro-scale study, that is, transport properties in cement paste. The REV size of cement paste for transport properties is found to be 100×100×100 ?m3. The water permeability, moisture and chloride diffusivity in cement paste are simulated using the PermLBS module, SCMPLBS module, and DiffLBS, respectively. The effects of w/c ratio, time, chloride binding, degree of water saturation, and chloride concentration on the chloride diffusivity in cement paste have been studied in a quantitative manner. The simulation results are verified with the experimental data from literature. The predictions by the micro-scale solver agree quite well with the experimental data. Chapter 7 focus on the meso-scale study, that is, transport properties in mortar and concrete. The diameter of the used sands ranges from 0.125 to 4 mm. The REV size of mortar for transport properties is found to be 8×8×8 mm3. Chloride diffusivities in mortar and concrete are estimated using the meso-scale solver. The effects of ITZ, aggregate content and shape on the chloride diffusivity in mortar are quantitatively investigated. The simulation results of chloride diffusivities in cement-based materials at meso-scale show a good agreement with the experimental data from literature. Chloride-induced corrosion of steel rebars in the concrete is considered to be one of the major causes of deterioration in reinforced concrete structures. In order to demonstrate the potential applications of the developed multiscale modelling scheme to predict the service life of reinforced concrete structures, a case study is presented in Chapter 7. It is shown that the estimated transport properties in cement-based materials by the developed multiscale modelling scheme can be directly used as input for the service life prediction of reinforced concrete structures.