Stability of building structures using NLFEA including second order effects : Applications with the software ATENA-GiD

Abstract: In Sweden, prefabrication in the construction industry is a very common system. In prefabricated structures, there are various kinds of connections and generally many hinged connections. Therefore, it is of high interest to study the behavior of the prefabricated buildings in relation to the connections. Two main theories that can be used to evaluate our structures are the non-linear finite element analysis (NLFEA) and the second order effects. These two theories help to study the real behavior of the structures. Software that can apply these two theories is the ATENA-GiD. Therefore, it was very interesting to explore this software and find out how it can be used for prefabricated buildings. For this reason, a simplified ten-storey building was studied which had been also used in Lindwall’s and Wester’s (2016) master thesis. Various cases were chosen for this building. The building was studied either with monolithic walls or with vertical slip connection between the prefabricated walls. Also, the hollow core (HC) slabs were modelled in two ways and the supports of the columns were modelled in three ways. Even when each slab was modelled with two materials (solid concrete for the upper and lower layer with smeared reinforcement for the lower layer and modified concrete in order to compensate for the voids), the increase in the time for analysis was not important. Also, the cracking was not extensive thanks to the relatively low wind load. The slip joint between the walls didn’t lead to the different displacement between the walls. It was found that the hinged support in the columns was modelled by using the upwind line of the base. In non-linear analysis cases, the model of the structure gave 20% larger displacements than in linear analysis cases which means that the cracking had an important influence on the second order effects despite that the cracking was not extensive. The reduction of the stiffness of the walls to 40% accounting for cracked sections in additional linear analyses to according to the Eurocode 2 was a conservative choice for this building. The most remarkable result was that when the supports of the columns were modelled with hinges there was smaller displacement than when the supports were fixed. The reason for this was the negative displacements due to the buckling of the columns of the first storey. Moreover, for the linear cases, there was a much smaller displacement (5mm) than Lindwall and Wester (2016) (17mm) and this was possibly due to the way that the hinges were modelled in my building. It is interesting to investigate other ways to model the hinges in ATENA-GiD as well as other ways to model the connections between the HC slabs and the perimetric beams. Also, the effect of creep through the modulus of elasticity could be analyzed. Finally, it would be interesting to include foundations in the model to study the possible need for anchorages.