Seismic Performance of Prefabricated Steel Frame Structures with Buckling-Restrained Braces

Abstract

The concept of “prefabrication” has been increasingly applied in the civil construction industry. While satisfying the requirements for the rapid development of advanced modular constructions, prefabrication may not be an appropriate choice in earthquake regions due to its poor seismic performance. On the other hand, as one of the most favoured options in seismic design, steel structures also face many challenges. Most of the steel frames used in strong seismic areas have field-welded moment-resisting connections. This is despite the fact that on-field welding of beam-column connections lacks construction quality assurance and is prone to industrial accidents. The brittle fractures in welded steel beam-column connections caused by earthquakes can also lead to structural collapse. Further, the costs of repairing or replacing the earthquake-damaged structures can be exorbitant. To satisfy both seismic and prefabrication requirements, this PhD dissertation recommends a prefabricated steel-frame system with bolt-connected beams and columns incorporating buckling restrained braces (BRBs). In this system, the welded rigid connections are replaced by bolted connections which only require on-site assembly by bolting. The potential damage to the main structural members caused by earthquake ground motions can be effectively reduced. This is because the BRBs are introduced to resist most of the lateral seismic loads whereas the bolt-connected beams and columns are designed to carry gravity loads only and hence they remain elastic. The focus of this dissertation is on the seismic performance of the proposed bolted-frame system with BRBs. To this end, a methodology for the numerical modelling of the structural behaviour of the system using OpenSees is developed. Its effectiveness is validated against existing experimental results in published literature. An experimental study including the tests of three 2-storey frames has also been undertaken to investigate their dynamic responses under cyclic loadings. The experiment results are discussed and compared with numerical predictions in terms of hysteretic behaviour, energy-dissipating capacity, and stiffness degradation. In addition, a parametric study is carried out on two groups of models namely the R- and P-Groups each of which comprises 9-storey and 5-storey frames. The frames in the R-Group are with moment-resisting beam-column-brace connections; those in the P-Group are with non-moment-resisting ones. The emphasis of the parametric study is on the impact of the beam-column-brace connection rigidity and the BRB area distributions on the seismic resistance capacity of the system. Furthermore, a fragility analysis is conducted to fully understand the structural demand caused by various levels of ground shaking. For this purpose, 12 prototype frames with three different storey heights and various numbers of non-moment resisting bays have been designed and analysed. This analytical and experimental PhD research has led to the conclusion that the proposed bolted-frame system with BRBs is able to provide a reliable performance under seismic actions. In general, the system performs well in energy dissipation under severe earthquakes, especially if the BRBFs are constructed with moment-resisting beam-column-brace connections. In terms of the stiffness ratio (Sr), the ideal value is found to be between 3 and 5 regardless of the building height. For the design distribution of BRB areas along the storey height, the most uniform patterns of the inter-storey drift ratios are produced when the said areas are distributed proportionally to the storey shear force. Further, in terms of fragility, with the same building height, the probabilities of exceeding a given performance level are always higher with more numbers of NMRBs, and with the same numbers of NMRBs, the taller buildings have less likelihood to collapse. Finally, in regard to collapse prevention, the proposed system is only recommended to use with no more than 3 NMRBs.