The micro-structure evolution and macro-property development of cement-based composites are affected by the wall effect. For the cement paste, the formation of the interfacial transition zone (ITZ) is related to wall effect, i.e., the lower density of cement particles near aggregate induces the formation of ITZ. In addition, the aggregate density is low near the formwork during the concrete placement and molding, which results in a difference of properties between the wall effect zone and the material interior. Characterizing the influences of aggregate or cement particle shape and size-polydispersity on the thickness of wall effect layer and further on the volume fraction distribution of particles is important for controlling the performance of cement-based composite. Compared with theoretical and numerical simulation methods, the experimental method, such as concrete section cutting is time-consuming and labor-intensive. With the development of computer technology, it provided a technical basis for the study of wall effects in concrete through numerical simulation. In general, cement-based composite can be regarded as a particle packing system. However, most of the existing numerical and theoretical studies employs simple particle, such as circles, ellipses and spheres to model the aggregate or cement particles, which makes it difficult to accurately analyze the role of particle shape and size-polydispersity on the wall effect. It is thus of great significance to construct a complex particle packing system and develop a micro-structure characterization algorithm applicable to the particle packing system to investigate the influence of the particle shape and size-polydispersity on the wall effect.

This work used variable shaped ovoidal particles to construct aggregate and cement particles. A contact detection algorithm for ovoidal particles was proposed based on a geometric potential concept, and ovoidal particle packing systems were generated to simulate the packing state of aggregate and cement particles. Moreover, the cross-section analysis algorithm applicable to the ovoid particle packing system was proposed. Based on the cross-section analysis algorithm and the stereology unbiased estimation method, the influences of particle shape and size-polydispersity on the thickness of the wall effect layer were analyzed and a quantitative relationship between the thickness of the wall effect layer and particle shape and size-polydispersity was established. A prediction formula was proposed to determine the particle volume fraction distribution at different section planes considering the wall effect, and the accuracy of the prediction formula was verified via comparing it with the results of the existing studies. Finally, the modulus of elasticity of concrete at different locations was calculated by the quantitative model of wall effect layer thickness and particle volume fraction distribution based on the differential effective medium theory.

For monoshaped-monosized particle packing system, the relationship between volume fraction of particle (V_V) and distance from boundary plane to cross-section plane (h) is obtained. The results show that the V_V-h curves for different shaped particle packing systems are basically similar, indicating that the influence of particle shape on the thickness of wall effect layer T_P can be ignored.

For monoshaped-binarysized particle packing systems, the V_V-h curves move in the direction of increasing distance as the proportion of larger size particles increases. The results indicate that the thickness of wall effect layer T_P enlarges with increasing the proportion of larger size particles. For monoshaped-polysized particle packing system, the thickness of wall effect layer T_P enlarges with the increase of average equivalent diameter of particles.

Based on the obtained results, the prediction formulas are proposed to determine the wall effect layer thickness and particle volume fraction distribution, and the reliability of prediction formulas is verified via comparing the predicted data with the results from this work and other publications.

The relationship between elastic modulus E^MT and distance h is similar to that between volume fraction of particle (V_V) and distance h. The variation of E^MT-h curves can be divided into the increasing and stabilizing parts. The results indicate that the elastic modulus E^MT is small near the boundary plane, which is induced at a low particle volume fraction.

This work could provide a guidance for the optimization of the microfine structure of concrete and the regulation of macroscopic properties, and the following conclusions were summarized: 1) The influence of particle shape on the wall effect was not dominant; 2) The thickness of wall effect layer T_P increased with enlarging the average diameter of aggregate, and the thickness of wall effect layer T_P decreased with enlarging the aggregate volume fraction; 3) The relationship between elastic modulus E^MT and distance h was similar to that between volume fraction of particle (V_V) and distance h.