J.G.M. van Mier

Experimental and numerical analysis of micromechanisms of fracture of cement-based composites

Abstract In this paper experimental evidence of fracture toughening of concrete and mortar through a mechanism called crack face bridging is presented. The classical explanation for softening of concrete, viz. the formation of a zone of discontinuous microcracking ahead of a continuous macrocrack seems only partially true. Instead, crack face bridging in the wake of the macrocrack tip seems a physically sounder explanation. The crack face bridges are flexural ligaments between ovelapping crack tips. The failure of the flexural ligaments occurs in a stable and controlled manner because the two overlapping crack tips shield each other. The cohesive stress over the macrocrack is directly related to the size of the crack face bridges, which depends on the heterogeneity of the material. The typical failure mechanism can be simulated using a simple numerical lattice model. First the grain structure of the material is generated either by manual methods or by adopting a random generator. Secondly a tringular lattice of brittle breaking beam elements is projected on the grain structure. Aggregate, matrix and bond properties are assigned to the lattice elements at the respective locations, and a simple algorithm allows for crack growth simulation. The main conclusion is that the crack patterns and the associated load-deformation response are largely governed by the properties of the constituents. The bond between aggregates and matrix is the weakest link in the system, and variation of this parameter leads to profoundly different crack patterns.

3D lattice type fracture model for concrete

A 3D beam lattice model is used for simulating fracture processes in concrete, which is schematized as a three-phase material (aggregate, interfacial transition zone and matrix). Numerical experiments are conducted varying the particle density. The obtained results show that, with increasing particle density, the peak-load decreases and the response is more ductile. This appears to be related to the amount of de-bonding. As a matter of fact, when the strength of the interface transition zone is set equal to the strength of the matrix, neither the peak-load nor the ductility of the lattice response are influenced by the particle density.

Strain-softening of concrete in uniaxial compression

0025-5432/97

Influence of microstructure of concrete on size/scale effects in tensile fracture

Abstract Size/scale effects on the fracture of concrete subjected to uniaxial tension are studied by means of analyses with the Delft lattice beam model and compared to recent experimental results. Using a simple local elastic–purely brittle material global softening behaviour is calculated. The effect of deterministic and statistical contributions to size effect is studied by implementing different degrees of heterogeneity to the lattices. They vary from ‘homogeneous’ regular triangular lattices to lattices with randomly varying beam length. Computer generated particle overlays are used to improve resemblance to real concretes. Trends in size effect on nominal strength and fracture energy are in close agreement with the experiments. The type of approach can be used for tuning macroscopic size/scale laws for concrete and related materials.

Mode I fracture of concrete: Discontinuous crack growth and crack interface grain bridging

Abstract In the present paper results are presented of vacuum impregnation tests using single-edge-notched concrete plates subjected to uniform boundary displacement. A realistic fracture model for cement-based composites should be based on correct micro-structural observations of cracking and localization. Macroscopic structural effects such as non-uniform opening and non-uniform drying out have a pronounced influence on the specimen behaviour. The results of the vacuum impregnation tests reveal that the fracturing of the specimens is a three-dimensional growth process. Cracking starts from the outer surfaces of a specimen, most likely due to non-uniform drying out and the related tensile eigenstresses near the specimens surface. The macrocracks are highly discontinuous cracks with debonding near larger aggregates and intact material bridges between them. The load carrying capacity of a tensile specimen for average crack openings larger than 50 μm can be explained from distributed crack interface grain bridging. The failure of the grain bridges is a process involving bending and frictional pull-out. Next to this an ‘intact core’ was observed in some specimens loaded up to average axial deformations between 50 and 100 μm. The effects of maximum aggregate size on softening was significant for the range of specimen sizes and aggregate sizes tested. The increased post-peak load carrying capacity for coarse grained mixes can be explained from the larger size (and thus the increased flexural capacity) of the crack interface grain bridges.

How to study drying shrinkage microcracking in cement-based materials using optical and scanning electron microscopy?

Abstract Three different ‘destructive’ microscopy methods were tested on their ability to show drying shrinkage microcracks on a specimen cross-section. The first two were methods in which the microcracks were impregnated with a fluorescent epoxy and examined with fluorescence microscopy. In one method, the impregnation was applied before making the cross-section and in the other after making the cross-section. In the third method, the sample was kept wet constantly and examined in an environmental scanning electron microscope (ESEM). It was concluded that the method in which the dried specimen was impregnated before making the cross-section was the most reliable method to record drying shrinkage microcracks. With this method, it was possible to impregnate the complete drying shrinkage microcrack pattern in the studied cement-based materials from the surface, and there was no risk of recording microcracks introduced by sample preparation.

Experimental investigation of concrete fracture under uniaxial compression

Localization of deformations has been investigated in a series of displacement controlled uniaxial compression experiments. Of main interest are the effects of specimen slenderness and friction between loading platen and specimen. Both effects have a direct influence on the development of localized fracture zones in the specimen. The results indicate that the use of a double layer of teflon with an intermediate layer of grease yields size-independent results as far as the pre-peak stress–strain behaviour and the peak strength are concerned. However, in terms of stress and strain, a significant influence of both the specimen slenderness and the amount of boundary restraint has been observed in the post-peak regime. It is found that the post-peak curves become almost completely identical when they are plotted in terms of nominal stress and post-peak displacement. For any type of loading platen used, the post-peak relative stress-displacement curves are found to be independent of the specimen height. Furthermore, since during post-peak localization relative sliding and movements of larger parts of the specimen are observed, the definition of a unique Poissons ratio is virtually impossible.

Effect of strain gradients on the size effect of concrete in uniaxial tension

A series of uniaxial tension experiments has been conducted to investigate the size effect on strength and fracture energy of quasi brittle materials like concrete and sandstone. This paper focuses on the results of the concrete tests, and specifically deals with the variation of the nominal strength for specimens of six different sizes in a scale range of 1:32. It was found that under given experimental conditions, the nominal strength strongly depended on the specimen size. More important however, is the fact that most of this size effect could be attributed to strain gradients which were present in the cross section of the specimens. These strain gradients were caused by the specimen shape, load eccentricity and material inhomogeneity. Through a combination of experimental data and a simple linear elastic analysis, the importance of strain gradients with respect to the ultimate load level could be visualized. This leads to the conclusion that studying a material size effect is not possible without taking into account structural stress/strain gradients.

Lattice model evaluation of progressive failure in disordered particle composites

Abstract Results from experimental and numerical investigation on the fractal properties of particle composites are presented. Splitting tests on concrete were carried out and the fractal dimension of the crack patterns detected on microscope images was measured. Numerical simulations of the tests were performed by means of a lattice model, and non-integer dimensions were measured on the predicted lattice damage patterns. The ability of the model to reproduce realistic statistical interactions and self-organization in the propagation of the cracks is discussed. In particular, an increasing Hausdorff dimension was measured during damage development in the lattice network. It is concluded that, next to the stable crack growth provided by fractality in disordered materials, an optimal choice of the percentages of weak and strong microstructural elements, together with their particle-like distribution, may lead to improved mechanical performance of the considered materials.

Crack growth mechanisms in four different concretes : Microscopic observations and fractal analysis

Splitting tests have been carried out for evaluating damage in four types of concrete. The concretes contained either river gravel with maximum aggregate size equal to 2 and 16 mm, phosphorous-slag aggregates, and Lytag aggregates. Different failure behaviours have been revealed, mainly depending on the aggregate and interface characteristics. The microscopic cracking mechanisms strongly affect strength, toughness, and ductility. In addition to mechanical measurements, a high resolution optical microscope was adopted for detecting the crack growth. Fractal dimensions of the crack lips and of the complex damage patterns were computed. No simple relation exists between fracture energy and fractal dimension, but relative differences in the material structure affect the value of the fractal dimension. This confirmed the peculiar different behaviours encountered in the tests.