In spite of very long research activities with the first tests conducted by Mörsch in the 1920th, the design for shear of concrete members without transverse reinforcement is still not cleared. This may be demonstrated in the fact that most shear design procedures are not based on mechanical models but on empirical equations which show very big uncertainties. Thus very high safety factors have to be used. The problem of shear design exists primarily for slabs. In Germany numerous old concrete bridges show a lack of safety in transverse direction since the year 2001, when a new DIN-code was introduced. Furthermore the shear design principle of haunched concrete structures, as for example, cantilever slab of bridges, is hardly cleared, up to now.
This research work aims to improve the understanding of the shear behaviour, to identify the significant shear resistance mechanisms and finally to develop new shear design models for practical use which is valid for straight and haunched concrete beams without stirrups.
The essential, partially very different approaches known from the literature are analysed first and their accuracies are checked by a comparison with the test values from a shear database. Considerable differences appear between the arithmetic results and the test values.
To study the behaviour of concrete members under high shear loads an extensive test program with 18 reinforced concrete beams without web reinforcement of different shear slenderness and inclination of compression chord was conducted. These experiments focused on haunched beams. It appears that the existing approaches describe the influence of an inclined compression chord on the shear capacity only in an insufficient manner. In some cases the design according to the DIN- or EC-Code results in unsafe values.
To get a better understanding of the load-bearing mechanisms all tests are simulated by means of non-linear Finite Elements analysis with the programme ABAQUS. Very good agreement appears between the test results and those of FE-analysis including the load-deflection curves, the load bearing capacity and the crack patterns. Thus the developed numerical model can simulate the behaviour of reinforced concrete beams up to the ultimate limit state. The FE-analysis demonstrates that the shear forces are transferred in uncracked compression zone mainly. Therefore crack friction and dowel action plays no significant role on shear bearing capacity at the ultimate limit state. As a result of the FE-analysis and the test program, a shear resistance action of uncracked concrete part in the tension zone is firstly introduced to be one of the two main shear bearing actions of concrete structures without stirrups.
Based on the theoretical and experimental investigations an easy analytic model is developed for the estimation of the shear capacity of straight and haunched reinforced concrete beams without stirrups. It shows more accurate results than the known approaches.