Abstract (EN):
Advances in computer technology allow nowadays the use of powerful computational models to describe the non-linear structural behavior of reinforced concrete (RC) structures. However their utilization for structural analysis and design is not so easy to be combined with the partial safety factors criteria presented in civil engineering international codes. Trying to minimize this type of difficulties, it is proposed a method for safety verification of RC structures based on a probabilistic approach. This method consists in the application of non-linear structural numerical models and simulation methods. In order to reduce computational time consuming the Latin Hypercube sampling method was adopted, providing a constrained sampling scheme instead of general random sampling like Monte Carlo method. The proposed methodology permits to calculate the probability of failure of RC structures, to evaluate the accuracy of any design criteria and, in particular, the accuracy of simplified structural design rules, like those proposed in civil engineering codes.
Language:
English
Type (Professor's evaluation):
Scientific