Crack widths in HSC

Abstract

This master thesis covers the research into cracking behaviour of slender high strength concrete cantilevering balconies. The research emerged from Pieters Bouwtechniek Delft and Hi-Con Denmark. They cooperatively designed very slender balconies in ultra-high performance fibre reinforced concrete. Other engineering firms tend to recreate these balconies in HSC and insecurities emerged around the cracking behaviour and crack width prediction for these slender HSC cantilevering balconies. In addition to this research an informative report, also functioning as a literature study, on the connection of prefabricated concrete balconies is produced. This report is separately attached.

The research is step wisely conducted starting with a simple fully clamped cantilevering slab. For this slab the cross sectional height, reinforcement diameter and reinforcement spacing are varied to investigate their influence on analytical crack width predictions. First for all variants an analytical design and analysis process is executed, followed by a numerical analysis with DIANA FEA and a comparison of the results. The most important observation is that for a cross sectional height of 120 mm or smaller reinforcement bars are located outside the effective area, making the analytical method unsuitable. Furthermore, a big discrepancy between the predictions of the different analytical models is observed, indicating an unreliability of these methods.

In two steps the fully clamped balcony is transformed into a Hi-con shaped balcony executed in HSC. It appeared that in light of detailing rules from Eurocode 2 an exact reproduction is impossible, but the concept could be reproduced in a less slender way. Furthermore, by comparing analytical and numerical design results for two different balcony designs it is found that the accuracy of the analytical crack width prediction depends on geometric disturbances. In case a geometric disturbance is present in a slender area loaded in tension peak stress concentrations occur, which negatively influence the reliability of the analytical crack width prediction. In case the area is less slender, the effect is less pronounced and the conservative characteristics of the analytical method outweigh the influence of the concentrated peak stresses.

When summarizing, it appears that specific care should be taken when analytically predicting crack widths in slender balconies because it might appear that the reinforcement is not located in the effective area. Furthermore, the more slender the structures become, the bigger the influence of a geometric disturbance can be, increasing the risk of an underestimation of the occurring crack widths because peak stress concentrations are analytically not accounted for.