Concrete has many advantageous properties as regards resistance to fire. It is non-flammable and it has a low thermal conductivity. However, concrete structures, which are not designed for resistance against fire, show significant damage after heating. In particular, the explosive flaking with the consequence of weakening the reinforced concrete cross-section and exposing the steel reinforcement to the temperatures higher than critical temperature of reinforcement. There are only a few possible measures of preventing or mitigating the effects temperature load used. Ways of protection can be divided into two systems: active and passive. Active systems are designed to ensure the greatest possible reduction of temperatures the concrete is exposed to. Passive systems directly resist to high temperatures and fire. Design of composition of concrete with the aim of higher resistance to exposition to high temperatures belongs among the passive systems. A part of the work focuses on summary searches of the problems of concrete and reinforced concrete structures exposed to high temperatures and fire. The goal of the work is defining requirements for cement matrix based composite material and its design ensuring the highest possible resistance to high temperatures or direct fire.

The thesis is focused on a test on determination of the fracture-mechanical parameters of quasi-brittle materials, especially concrete. What is referred to as the wedge-splitting test is considered, for which a variety of shapes of notched specimen can be used. This work is exclusively focused on the cylinder-shaped specimen of diameter 150 mm and breadth of 100 mm. The test is performed virtually using Atena 2D FEM software. Progress of failure is observed during loading of the specimen for various notch lengths. The amount of energy released for the development the failure outside of the tested cross-sectional area (weakened by the notch) is quantified and the size of the fracture process zone is investigated. The described analysis is performed for several material sets witch differ in cohesive properties of the quasi-brittle material expressed via the so-called characteristic length. Suitable proportions of the test specimen are sought, in order to avoid the failure and thus also the energy dissipation outside of the specimen ligament area during the experimental tests, which shall lead to more accurate estimates of fracture-mechanical parameters of the tested material.