Abstract |
: |
Two-blade bridge piers represent today an innovative solution to optimize the structural behavior of long viaducts subjected to seismic actions (Biondini 2001, Calzona & Bontempi 2001). In fact, a proper design of the structure provides a correct transmission of vertical loads to the foundation structures and the possibility to allow horizontal displacements without the use of kinematics mechanisms. Due to the particular configuration of the pier (a rigid box section in the bottom part and two flexible blades at the top connected with the viaduct deck), the structural behavior can be achieved only using refined analysis formulations and effective numerical tools. In particular, during the nonlinear response, the shear effects are very important and can govern the structural behavior. Nonlinear models have to be able to consider the shear damaging with sufficient accuracy otherwise, the numerical response would be affected by heavy errors. In addition, uncertainties on material characteristics can affect the numerical response. To handle these important uncertainties involved in the problem, the use of probabilistic or fuzzy approaches can be suitable. In this paper, to show the influence of the degradation on the structural response a new and a damaged two-blade pier will be analyzed. Probabilistic evaluation of the material properties are considered to improve the reliability of the numerical analyses. The material properties on the numerical model are not considered in deterministic way, but having a probabilistic distribution. Physical consideration has suggested to model the material properties with a Lognormal distribution. To reach a significant value of the solution in the probabilistic approach, 30 piers are considered with different mechanical properties. To study the influence of the position of a generic deterioration, one has introduced in the model a portion of the pier deteriorated having 4 m of height. Clearly, the quota of the deteriorated section (Figure Presented) influences the global response of the pier. Therefore, one has considered 19 different quotas in order to study the structural sensitivity at the deterioration. For each different quota, 30 different piers with initial mechanical properties, assumed as stochastic random variables having a Log normal distribution, were considered. To summarize the results, it is possible to define a reduction factor like: r = (f)/(f und) where (f) is the expected value of f greatness, for the 30 evaluations performed, and (fund) is the expected value of f for the 30 not degraded piers. In this way, the reduction factor r is 1 if the pier is not damaged, and assumes values among 0 and 1 in the other cases. The left curve in the Figure 1, represent the reduction factor of the displacement while the right curve represent the reduction factor of the strength. It is possible to note that the reduction factor increase quickly when the deteriorated portion is close to the section (*). Therefore, it is possible divide the pier in two zones. The zone A) in which the structural behavior is strongly influenced from a deterioration and the zone B) where the deterioration has not effect on the nonlinear behavior of the structure. |