Projeto semi-probabilístico de vigas de concreto armado com modo de falha por cisalhamento

Abstract

The shear strength in reinforced concrete beams (RC) has been the subject of intense research for more than 100 years. The size effect on the shear strength of these beams is also known since long. However, there is a lack of consensus among researchers reflected by the different models prescribed by NBR 6118 (2003) and the ACI 318 (2008) as well as the non-explicit inclusion of the size effect in these formulations. Within this scenario, the objectives of this work are: (i) to develop a model to estimate the strength of RC beams subjected to shear, (ii) to obtain the statistics associated to the model error, (iii) to assess the reliability of beams designed according to NBR 6118 and ACI 318, and (iv) from the resulting reliability indexes, to propose a design formulation for shear consistent with the recommendations of NBR 6118. The formulation for strength estimation was developed by regression analysis of test results obtained from the literature. The selection of data was based on the similarity of the values for the relevant parameters and the number of data-points for different height ranges. Unlike many code formulations considered in this work, the proposed model is not biased with respect to the variables of interest (beam height, compressive strength of concrete, etc). The reliability evaluation, via Monte Carlo simulation, uses the regression model to calculate the shear capacity of the beams and incorporates the random variable model error. The reliability indexes obtained show little uniformity (1.68 2.66, NBR 6118 and 2.07 2.89, ACI 318), inadequate minimum values, and are also reasonably lower than those reported by other authors. It was observed that such discrepancies have their origin in the value of the coefficient of variation found for the model error (23%), much higher than that used in other studies (10%). A new model for shear design of RC beams is proposed for the NBR 6118, resulting in greater uniformity and higher reliability indexes.