Crack width in Tunnels

Abstract

When engineers at Tunnel Engineering Consultants (TEC) used the Eurocode 2 (EC2) for the verification of flexural cracks in cross sections of immersed tunnel elements, they found out that there was more reinforcement needed for crack width limitation than for tensile strength of the reinforced concrete. This raised questions, namely whether the extra reinforcement is needed, how accurate the EC2 calculates the crack width and what the influence is of the construction height on the crack width?

Are the current rules in the Eurocode 2 for the calculation of flexural cracks too conservative for large thicknesses that are applied in the concrete lining of immersed tunnels?

In this thesis the crack width calculation of the EC2 has been compared with other codes around the world. When these codes are compared with each other it seemed to be that the EC2 is not conservative in comparison with the other codes, however it does not directly mean that this is true.
To check whether the codes, especially the EC2, are conservative, the codes should be compared with laboratory experiments. Since there is little data of experiments available where the cracks are carefully measured, in particular beams with large thicknesses, the finite element program DIANA is used to gain more data sets. A DIANA model is validated with three different experiments with each a different height.

When the EC2 and the DIANA results are compared, a few differences are found. The predicted crack widths in the EC2 are larger than that of the DIANA results. In the EC2 the crack width is the multiplication of the crack strain and crack spacing. If these values are compared, the EC2 gives a smaller crack strain than DIANA and therefore the crack spacing gives an even larger difference than the crack width. Next to that, the influence of the parameters in the EC2 is analysed. In both the EC2 and the DIANA results the cover is the most important parameter for the crack spacing and the steel stress of the reinforcement is the most important parameter for the strain. In all the results there is almost no effect of the construction height visible, except in the crack strain of the DIANA results when the cover is relatively large. A decrease of strain with an increase of height has been detected.

The differences in crack strain can be explained by the fact that the strain in the EC2 is calculated at reinforcement height, but the strain at the outer fibre is needed. In axial loaded cases these strains are the same, but in cases where the cracks appear due to bending moments, which is the case in this thesis, the strain is bigger at outer fibre. Next to that it seemed that influence factors used in the crack spacing calculation are overestimated, but no clear reason is found for this result.