1. Introduction
Numerous structures are recently required continuous use even after large earthquakes. An efficient technique is to employ elasto-plastic dampers, such as buckling-restrained braces (BRBs), which act as fuses. Although response spectrum analysis (RSA) has a good time efficiency for damper design, relative to non-linear response history analysis (NLRHA), the conventional methods are generally limited to specific structures that are easily translated into simple models. With a suitable RSA method, computational optimization of damped structures will be also important, considering recent trends towards the use of algorithmic modelling, scripting and cloud-based high performance computing servers. Thus, this paper presents a damper design routine for highly indeterminate 3D structures utilizing computational optimization and response spectrum analysis, which has been extended to incorporate non-proportional damping.
2. Generalized Response Spectrum Analysis and computational work flow
The proposed routine (GRSA) is presented, and the accuracy is verified in the comparison with THA.
3. Damper layout or distribution optimization for the lattice tower
A series of optimizations, including damper layout or distribution, is performed for the lattice tower in order to verify the efficiency of the proposed design routine, and to analyze the efficiency of the existing retrofit design.
4. Damper distribution optimization for the conventional BRBFs
A multi objective damper distribution optimization is performed in order to verify whether the well-known performance curve can be found in a detailed BRBFs. Additionally, the optimization results are compared with that of the conventional design method.
5. Layout optimization for a concentric braced frame skin structure and the practical optimization conditions
A series of detailed layout optimizations, including BRB replacements of the concentric braces, BRB replacements of the first story columns or the concentric brace removals, is performed in order to investigate both the optimal layout tendencies and the practical optimization conditions.
6. Conclusions
In summary, the following results were obtained:
1) GRSA can be used for direct RSA of 3D structural analysis models including dampers of various types, with buckling-restrained braces used in this study, and the accuracy is verified in a series of comparison studies with NLRHAs. In the study of the systems including 810 degree-of-freedoms, NLRHAs (7,000 steps) take 15 to 30 min., GRSAs take 5~10 s, which indicates GRSA has a good time efficiency compared with NLRHA.
2) While improved layouts and distributions were identified for certain constraint and target objectives, the existing design for a particular lattice tower was found to be close to the optimal solution for minimizing the buckling utilization. While the layout and distribution trends are usefully indicated by the optimal solutions, the results are dependent on the input waves, suggesting that there may be no versatile optimum solution to every possible scenario.
3) Even for the multi degree-of-freedoms models, the well-known performance curve related to the response reduction of passively-controlled buildings can be found, utilizing a multi objective distribution optimization method. While the full proportional distribution method of equivalent single degree of freedom is reasonable for low rise structures.
4) A single objective optimization may be an inadequate optimization condition for the concentric braced frame skin structures. A penalty is suitable for this situation in order to reduce both the story drifts and demand capacity ratios of member buckling. Nevertheless, BRB replacements of columns or concentric brace removals enable simple single objective optimization.
