Performance evaluation of a three story steel structure and retrofit with using fluid viscous dampers

After Northridge earthquake in California extensive damages at steel building connections were observed unexpectedly. SAC joint venture, funded by FEMA, was established after the Northridge earthquake to further investigate the possible causes of those unexpected damages. For the structures that have not been collapsed during the earthquake, it can be assumed that they have fulfilled their mission. However, after a very detailed investigation of the damages on steel structures, it was revealed that most of the welded beam-column connections were damaged beyond expected. The inspectors and engineers were very surprised after these findings as steel structures were believed to be invincible by the engineers before the earthquake. SAC joint venture hired different independent design firms to design three, nine and twenty story buildings according to the pre-Northridge and local codes in Los Angeles, Seattle, and Boston. After the designs were complete, extensive linear and nonlinear analysis were conducted to better understand the deficiencies of the damaged steel structures during the Northridge earthquake. Also about 120 full scale connections were tested to improve the steel moment connections and their load bearing capacities in the scope of that study. Three story Seattle structure designed in the scope of SAC studies is mainly investigated in this thesis. The performance of this structure is evaluated by using nonlinear pushover and nonlinear time history analysis on a 3D model. After the evaluation phase, the structure is retrofitted with using fluid viscous dampers. Up to date there is no application of the fluid viscous dampers in buildings' design or their retrofitting design in Turkey. But this technology is being widely used especially in United States of America and Japan and their design procedures are well defined in design codes. Dampers can mainly be categorized as displacement dependent devices or velocity dependent devices. However dampers that include the properties of both displacement and velocity dependent devices are also available (visco-elastic dampers). All damper types has their own advantages and disadvantages but the primary advantage of fluid viscous dampers over conventional retrofit methods is, as these devices are velocity dependent, their reaction forces are out-of-phase with displacement. Performance concept for structures are defined firstly with 2007 Turkish Seismic Code in Turkey. However, performance evaluation procedures and performance acceptance criterias are only defined for reinforced concrete structures. As performance evaluation procedures and performance target concepts form an important part of this study, performance concepts and definitions given in FEMA 356 and DBYBHY 2007 codes are also briefly explained. Procedures defined in ASCE 41-06 and FEMA 356 are mainly followed for performance evaluation of the investigated structure as no definitions exist for performance evaluation of steel structures in 2007 Turkish Seismic Code. When compared, it is seen that the definitions for different performance levels (Life Safety, Immediate Occupation etc.) are similar for investigated codes, however structure performance cannot be defined without the performance of nonstructural items in a structure for the ASCE 41 and FEMA 356 codes. As fluid viscous dampers will be used in the retrofit design of the building, a summary about fluid viscous dampers and damping on structures is also included. Damping can be described as decay with time in amplitude of a free vibration. The energy of a vibrating system is dissipated by various damping sources (opening and closing of micro cracks for concrete structures, interaction with nonstructural elements, friction in steel connections etc.). Contrary to conventional retrofit methods, fluid viscous dampers do not have a stiffening affect on the structure. Therefore, as the natural vibration period do not decrease, the earthquake forces acting on the structure do not change. Fluid viscous dampers introduce supplementary damping to the structure, which allows a serious amount of the earthquake input energy to be dissipated by added dampers. Viscous dampers can be defined by their damping coefficient and damping exponent. Dampers with damping coefficient equal to unity are named as linear dampers. Force output of linear dampers is proportional to velocity. If the damping exponent is not equal to unity, dampers are defined as nonlinear. Force output of nonlinear dampers are dependent on damping exponent and are not linearly proportional to velocity. In practice, it is possible to produce dampers with damping exponent ranging from 0.2 to 2. However, dampers with damping exponent between 0.3 and 1 are generally preferred for structures. The most important advantage of viscous dampers is, for linear dampers, the output forces of those devices are always out of phase with inertia forces of the structure. The maximum internal forces on the sections forming the structure are expected to occur at maximum drift for the investigated location, as linear viscous dampers provide their maximum output force at maximum relative velocity, those forces do not occur at the same time and therefore the possibility of additional retrofit (i.e. retrofit of foundations) on the structure is greatly reduced. In the scope of this study, the investigated structure is explained in detail including the analysis assumptions made. The studies carried out by Gupta and Krawinkler as well as FEMA 355 are mainly followed in this thesis as those studies are widely adopted for performance evaluation of steel structures in the current practice. Definitions given in ASCE 41-06 and FEMA 356 are used for steel column, beam and panel zone hinges for the nonlinear pushover and time history analysis. The possible affect of connections are not taken into account and were considered as fully restrained. A summary of pushover analysis and nonlinear time history analysis is also given including the steps and methods followed in this thesis. After the evaluation of the results obtained for the investigated structure, steps for determination of damper properties (damping constant, damping exponent etc.) are also summarized. The structure is assumed to be located at 1st degree earthquake zone with Z3 class soil according to 2007 Turkish Seismic Code. The performance of the structure is first evaluated with nonlinear pushover analysis and than the obtained results are compared with the results of three nonlinear time history analysis, where used acceleration records were matched to target spectrum. The obtained results for two different evaluation methods were mostly compatible with each other. The analysis results showed that the existing structure satisfies Life Safety performance level criterias for both directions considering the plastic rotation limitations given in ASCE 41-06 and FEMA 356. Even if there is no defined restrictions for structural performance taking into account of the drift ratios of the structures, analysis results showed that the drift ratios for the inspected structure exceed 0.03 for both principal directions it is decided that the existing structure should be retrofitted. When residual drift ratios are investigated from the results obtained from nonlinear time history analysis, it is seen that about %0.7 residual drift occurs at the structure. Decreasing the drift ratios of the structure below 0.02 is selected as the retrofit target. For preliminary analysis, as it is needed to decrease the structure's drift ratio by roughly 0.01, a damping ratio which will decrease the forces acting on the structure accordingly (about %20 of critical damping) is selected. After, damper properties and layout is selected considering the retrofit target. Four dampers is placed for each direction at each story of the structure to limit the maximum output forces of the dampers. The assumptions made at the preliminary design stage is than verified by nonlinear time history analysis. Time history analysis is carried out only for DBE level earthquake using three spectrum-matched records, therefore the maximum values obtained from the analysis are used. After the addition of the dampers, it is seen that the structure satisfies "Immediate Occupancy" performance level criterias for hinge rotations in both directions. It is also observed that the drift ratios of the structure are below 0.016. It has been seen that viscous dampers can be a very effective retrofit solution. With increased damping, the forces that the structure is expected to carry decreases and as the maximum damper output force is out of phase with inertia forces, no additional forces are introduced to the structural elements compared to the existing situation. This retrofit method can also be considered for reinforced concrete structures. The usage of steel structures in Turkey is increasing everyday, but the required parameters for performance evaluation of steel structures are not defined in DBYBHY 2007. It is seen from the carried out studies that, steel structures with low axial loads can go through extreme deformations without strength-degradation and can still satisfy Life Safety requirements in means of hinge rotations defined at ASCE 41 and FEMA.