The design shear resistance of an existing structure can be evaluated with analytical design procedures and numerical procedures provided by non-linear finite element analyses. The new fib Model Code 2010 proposes different calculation methods that fall into four different levels of approximation. As the level of approximation rises, so the complexity and the accuracy of the calculated shear resistance increases. Non-linear finite element analyses belong to the highest level of approximation, but although they are more and more becoming a customary tool in the daily design process, building codes do not provide guidance on how to perform these analyses.
This paper describes non-linear finite element analyses performed on prestressed beams, which underwent shear failure during experimental loading, in order to assess and criticize the finite element approaches. The aim of this work is to propose guidelines for numerical simulations in order to reduce model and user factors. The results obtained from the non-linear finite element analyses have been compared with the analytical results using different levels of approximation. The design shear resistance obtained with the highest level of approximation, level IV, derived from non-linear finite element analyses, turned out to be higher than the design shear resistance obtained with analytical procedures (levels I/II/III).