Examination of flexural crack width prediction in concrete: comparison of analytical and numerical models

The reliability of crack prediction methods in concrete design plays a role in the degree of confidence with which durability can be ensured. Bond failure between concrete and the embedded reinforcing steel exposes the steel surface within a crack path. This relative slip results from differential tensile strain between concrete and steel that allows harmful ions to reach and then react along the rebar length. A reliable prediction method should therefore account for the loss of bond in crack propagation. Researchers question the significance of the role played by surface cracks in structural deterioration, therefore casting doubt on the need for exhaustive crack analysis. The applicable fundamental theory of cracking, namely non-slip or slip determines the steel exposure and therefore the likelihood of reduced structural service life attributable to crack behaviour. However, while cracks originate at the surface of reinforcement through bond failure, simultaneously a cover distance away no cracks could appear on the concrete tension surface or they could be twice to ten times the crack width at the rebar level. Due to the heterogeneous composition of concrete, some commentaries state it impossible to accurately predict crack widths. Design standards therefore provide estimates of maximum crack widths to a degree of probability. This study examines the methods available for predicting cracks on the tension surface and the degree to which this is indicative of weakened bond between concrete and reinforcement. In this examination, it will be seen that concrete has ductility due to tension softening behaviour. The addition of steel to the tension area transforms the fracture process zone problem to the definition of a bond-slip relationship. Bond stresses generated at the rebar perimeter define the analytical relationships in design codes. These stresses control crack width at the tension surface. Results from the analytical code-based models are compared for increasing section depths, bar sizes and maximum spacings in the tension zone. A significant variation in the predicted maximum crack width is observed for deeper members. For very large concrete sections, the analytical models appear to provide unreasonable crack width values. The analytical equations in design codes focus on the bond relationship and ignore the size effect of concrete inherent in its microstructure. The concentration of flaws increases in larger members; hence size effect would play a greater role. Numerical modelling for crack prediction is therefore recommended for crack analysis in larger concrete members.

Grantor

Department of Civil Engineering

Year dc:date.issued

2020

Author dc:creator

• Fredericks, Brandon

Advisor dc:contributor.advisor

• Beushausen, Hans-Dieter

• Structural Engineering and Materials

Handle dc:identifier.uri

http://hdl.handle.net/11427/32675

OAI identifier oai:identifier

oai:open.uct.ac.za:11427/32675