For the dimensioning of concrete structures it is customary to determine the forces and stresses occurring in construction elements on the basis of the prescribed loads. In an iterative process, the final dimensions of the elements and reinforcement are then determined. This process becomes less simple if not only the external loads, but also the deformations imposed by shrinkage and temperature change are the cause of forces and stresses. The determination of stresses in the concrete becomes even more complicated if we not only want to know the distribution of forces and the associated stresses in the structures, but also the socalled eigenstresses, which are the result of non-linear temperature and shrinkage fields in concrete cross-sections. These eigenstresses can give rise to microcracks, often in the surface zone of concrete elements. These microcracks can jeopardize the durability of the concrete. With the methods for calculating the force distribution in a concrete construction, these eigenstresses remain out of sight. In current design practice, these eigenstresses are almost never taken into account. The assumption is that as a result of relaxation the eigenstresses largely disappear and will have no influence on the behavior and durability of concrete structures. This research examines whether this assumption is justified...

When designing a concrete structure, a structural engineer often assumes that it is initially free of stresses and cracks. Serviceability limit state analyses are therefore often carried out only taking into account the structure’s self weight and the other imposed loads. As a result, initial tensile stresses and (internal) cracking that can arise from, e.g., partially restrained shrinkage is not accounted for. This might imply that service life predictions can be too optimistic since micro-cracks can promote the penetration of chlorides, with the result of an accelerated corrosion of the reinforcement. The impact of restraint can be at both the structural level (plane sections must remain plane) and the material level (concrete is a non-homogenous material from components having different shrinkage behaviour and stiffness). In this contribution the results of the numerical analysis of eigenstresses and micro-cracking due to the shrinkage of the cement paste in concrete are highlighted. The two-dimensional Delft beam lattice model (BLM) is used to simulate concrete on meso-level as a three-phase material. Stress relaxation in the cement paste is calculated with the activation energy method. For the comparison of the calculated micro-cracking the results of laboratory tests on concrete are used by which the chloride penetration is measured under different restraint conditions with different levels of micro-cracking. The results of the numerical analyses with BLM specimens showed that micro-cracking can increase the diffusion coefficient of (uncracked) concrete with a factor 1.2–1.6, depending on the magnitude of stress relaxation. The residual tensile eigenstresses in concrete reduces due to micro-cracking and stress relaxation till values in the order of 0.5–1.0 MPa which are not taken into account in global calculations and can speed up (unexpected) cracking in the concrete.