A Damage Mechanics Based Random-Aggregate Mesoscale Model for Structural Concrete Analysis

• Main Author: Ni, Zhen
• Format: Dissertation
• Language: English
• Published: ProQuest Dissertations & Theses 01.01.2022
• Subjects: Aggregates; Cement; Concrete mixing; Construction; Crack initiation; Failure analysis; Load; Marketing; Mechanics; Medical imaging; Particle size; Probability distribution; Reinforced concrete; Statistics; Stress-strain curves; Tomography
• ISBN: 9798374487275

Conventional numerical solutions treat concrete as a homogeneous material, which may overlook the local cracking resulted from the heterogeneous nature of concrete. Such local cracking may further develop into catastrophic failure of the whole concrete structure if there is no sufficient reinforcement in the form of either steel bars or fibres. Concrete mainly consists of three phases, namely mortar, coarse aggregate, and the interfacial transition zone (ITZ) formed around the coarse aggregates. For normal concrete, the ITZ phase is usually the weakest link where damage initiates due to its weaker strength of the cement matrix. The distribution of the coarse aggregates therefore plays a significant role in the structural performance of normal concrete. The objective of this PhD study is to propose an innovative mesoscale model, which integrates the random distribution of the coarse aggregates and the damage mechanics of the ITZ and mortar phase, to simulate the realistic cracking of structural concrete under compressive and tensile/fracture failure.The proposed random-aggregate mesoscale model is firstly verified with the compressive test of a series of concrete cubes, whose mix proportion consists of the same mortar but different volumes of coarse aggregates. For each mix, the average response of different mesoscale models conforms with the experimentally measured stress-strain curve. The statistical distribution of the compressive strengths for each concrete mix is subsequently established using thirty mesoscale models with different random aggregates’ distribution, where both the Gaussian and Weibull distribution fit well with the rank probability.This research then adopts the proposed random-aggregate mesoscale model to study the fracture failure and the size effect of geometrically similar single-edge-notched beams (SENB) subject to three-point bend loading. The size effect of concrete can be classified as deterministic and statistical size effect. The deterministic size effect has been verified whereas there is seldom confirmation on the statistical size effect due to the lack of enough experimental data and reliable numerical methods. This study will demonstrate that the proposed random-aggregate mesoscale model can simulate the fracture failure of geometrically similar concrete specimens with the implementation of the correction factors, which eliminates the mesh sensitivity of concrete specimens with various dimensions. With enough different random distribution of the coarse aggregates, the proposed method can be further applied to determine the statistical size effect of concrete due to the random distribution of the coarse aggregates.Finally, this thesis applies the proposed method to simulate the performance of steelconcrete-steel (SCS) composite structures. It will identify the potential failure of the SCS composite beams, with and without the angle-connector reinforcement, due to different random distribution of the coarse aggregates.