Issue 50

S. R. Pereira et alii, Frattura ed Integrità Strutturale, 50 (2019) 242-250; DOI: 10.3221/IGF-ESIS.50.20 249 Regarding the predictions of the Eurocode 2, the strength of the column was very close to the one obtained experimentally in most tests, with only an underestimation of 28% in model E1. All other models presented less than 15% of deviation. This indicates that current standardized procedures are also applicable to elements with unusual cross-section. Figs. 10 and 11 present specimen E1 at the time of rupture in the experimental and numerical models. Figure 11 : Crack pattern in specimen E1 numerically obtained for the time of rupture. C ONCLUSIONS he behaviour of a precast reinforced concrete column for use in a modular housing project was evaluated according to different methodologies. To validate these methods, an experimental program was developed considering four different load eccentricities. The first proposed methodology was developed in accordance with the code requirements of the Brazilian ABNT NBR 6118:2014 standard [3] and is intended to describe the serviceability behaviour of structural reticulated reinforced concrete elements. The second methodology was based on finite element numerical models considering both physical and geometric nonlinearities. It was verified that, in general, both methodologies presented good correlation with experimental results regarding global serviceability behaviour. For the theoretical and numerical models of specimen E1, an overestimation of the initial stiffness in relation to the test behaviour was observed. From this point on, a sharp decrease in stiffness was noted. Both models behaved similarly to the experimental model until the serviceability limit state and, after this limit, a difference between values was observed. This was expected since the theoretical model does not present an adequate formulation for the evaluation of ultimate limit states. The numerical model of specimen E1 showed a strength about 10% lower than the experimental one and for specimen E2, this difference was 5% smaller. Differences of this order of magnitude can be considered appropriate mainly due to the complexities involved in this analysis, together with the uncertainties involved in the tests’ measurements. Numerical analyses for specimens E3 and E4, on the other hand, did not present such good correlation with experimental results, especially when ultimate limit state was analysed. While the numerical model for test E3 resulted in a high overestimation of the strength, model E4 presented a convergence problem with low ultimate load. Strength calculations according to Eurocode 2 led to good agreement with the experimental results, indicating that these standards may be adequately used to predict the load capacity of columns with unusual cross-sections. The models here proposed are feasible for the behaviour prediction of precast reinforced columns, and it is possible to extend the application of these methodologies to other practical situations such as bending of beams and slabs. Although the theoretical methodology proposed does not have such a high degree of sophistication compared with the Finite Element numerical model, a good correlation of results was observed, especially for lower loading levels expected in a serviceability situation. Therefore, this formulation is an alternative to the use of generic finite element commercial software, which require large initial investment and extensive operator training, in addition to presenting numerical convergence problems that are difficult to solve. A CKNOWLEDGEMENTS he authors would like to thank the funding agencies CAPES and CNPq for the support received and PRECON Industrial for the supply of materials, facilities and labour to conduct the tests. R EFERENCES [1] El Debs, M.K. (2000). Concreto Pré-Moldado: Fundamentos e Aplicações, São Paulo, Escola de Engenharia de São Carlos. T