iv ABSTRACT SEISMIC DESIGN OF COLD FORMED STEEL STRUCTURES IN RESIDENTIAL APPLICATIONS Uygar, Celaletdin M.Sc., Department of Civil Engineering Supervisor: Prof. Dr. &Ccedil / etin Yilmaz May 2005, 82 pages In this study, lateral load bearing capacities of cold formed steel framed wall panels are investigated. For this purpose lateral load bearing alternatives are analyzed numerically by computer models and results are compared with already done experimental studies and approved codes. In residential cold formed steel construction, walls are generally covered with cladding material like oriented strand board (OSB) or plywood on the exterior wall surface and these sheathed light gauge steel walls behave as shear walls with significant capacity. Oriented strand board is used in analytical models since OSB claddings are most commonly used in residential applications. The strength of shear walls depends on different parameters like screw spacing, strength of sheathing, size of fasteners used and aspect ratio. SAP2000 software is used for structural analysis of walls and joint force outputs are collected by Microsoft Excel. The yield strength of shear walls at which first screw connection reaches its shear capacity is calculated and load carrying capacity per meter length is found. The nonlinear analysis is also done by modeling the screw connections between OSB and frame as non-linear link and the nominal shear capacities of walls are calculated for different screw spacing combinations. The results are consistent with the values in shear wall design Guide and International Building Code 2003. The other lateral load bearing method is flat strap X-bracing on wall surfaces. Various parameters like wall frame section thickness, flat strap area, aspect ratio and bracing number are investigated and results are evaluated. The shear walls in which X-bracing and OSB sheathing used together are also analyzed and the results are compared with separate analyses.

Atualmente, os perfis de chapa dobrada têm encontrado intensa aplicação como elementos estruturais nas construções em aço e como consequência, estão sujeitos aos mais diversos tipos de carregamentos. Este trabalho apresenta os aspectos gerais do dimensionamento de perfis de chapa dobrada submetidos à flexão enquanto utilizados como vigas propriamente dito e, também, enquanto utilizados como terças. Analisam-se seções tipo U, tipo U enrijecido e tipo Z enrijecido. Os ensaios em perfis fletidos foram realizados sobre 3 diferentes condições de carregamento, com dois vãos distintos, em um total de sessenta e quatro perfis ensaiados. Para os perfis conectados às telhas de aço, os ensaios foram realizados em uma \"caixa de sucção\" desenvolvida especificamente para este projeto. Os perfis foram conectados às telhas pela mesa por meio de parafusos auto-brocantes, sendo realizados quinze ensaios. Os resultados dos ensaios são comparados com os resultados teóricos, resultados normatizados e resultados de alguns procedimentos aproximados. / Nowadays, the cold-formed steel members have intense application as structural elements in constructions and, in consequence, they are subjected to several types of loads. This work presents the general aspects of the study of cold-formed steel members subjected to flexion used as beams and, also, as purlins. Sections of type U, type C and type Z are analysed. The tests in beams were carried out under three different conditions of loads, with two distinct spans, in total of sixty four tests. For the purlins connected to roof system, the tests were in a \"suction box\", developed especially for this project. The purlins were connected to the roof system by the flange by using self-drilling screws, being carried out fifteen tests. The results of those tests are compared with the theoretical results, design specification provisions and results of some approximated procedures.

Terças de aço formadas a frio são muito utilizadas em sistemas de cobertura e fechamento. Para conferir continuidade entre tramos adjacentes e possibilitar um melhor aproveitamento de material, são empregadas ligações parafusadas por transpasse e com luva. O comportamento estrutural dessas ligações é muito dependente da sua configuração geométrica e do nível de carregamento. Com base em uma série de nove experimentos e análises teóricas, o comportamento estrutural dessas ligações foi investigado no tocante aos esforços resistentes e à rigidez. Observou-se que o modo de falha tem sempre uma componente distorcional significativa e uma resistência menor que a prevista pela metodologia atual de projeto. No caso do momento fletor resistente, a distribuição de tensões de flexão oblíqua é a que melhor se aproxima do comportamento real das terças, sendo possível a partir de uma análise de estabilidade elástica considerando tal comportamento, e, baseado no método da resistência direta e na curva de dimensionamento do modo distorcional, obter uma previsão segura do momento fletor resistente. Além disso, a rigidez da ligação, com relação aos deslocamentos verticais é bastante dependente do tipo de ligação, sendo que as terças com ligações por luva são mais deformáveis que terças sem ligações. Por outro lado, terças com ligações por transpasse são menos deformáveis que terças fisicamente contínuas. As parcelas de rigidez da ligação são apresentadas, e a componente de deformação do furo é a mais significativa. Uma expressão para deduzir a rigidez rotacional da ligação é apresentada, juntamente com um modelo de barras para modelagem da ligação. Os resultados teóricos se ajustam muito bem aos experimentais. Conclui-se em linhas gerais, que as ligações por transpasse podem ser consideradas plenamente satisfatórias do ponto de vista de comportamento estrutural, sendo que terças com essas ligações são mais resistentes e rígidas que terças com continuidade física. O mesmo não ocorre nas terças com ligações por luva, porém, dada a limitação no número de ensaios, tais ligações carecem mais estudos. / Cold-formed steel purlins are widely used in roofs and wall systems. The continuity of long runs of cold-formed steel purlins is guaranteed by sleeve and overlap bolted connections, which allows a better load distribution and material savings. The structural behavior of these connections is highly dependent on their geometric configuration and load level. The strength and stiffness of these connections were determined through a series of nine experiments, numerical and analytical analysis. Results have shown that distortional buckling has a major contribution on the stability analysis of the cold-formed steel purlin. To safely determine the flexural strength of the purlin, one must consider unrestrained bending stress distribution and distortional buckling. The connection type influences the rigidity of the purlin thus the vertical displacement. Purlins with overlapped connections are stiffer than physically continuous ones, although purlins with sleeved connections are not. The bearing deformation at the connection region is responsible for a reduction in the system stiffness. To account for the change in stiffness and determine the vertical displacement, an expression for the bearing stiffness and a beam model are proposed. The results of this model show good agreement with experimental data. The general conclusion of this research is: overlapped bolted connections enhance the structural response of the purlins when compared to physically continuous ones. The same does not hold true for purlins with sleeved connections.

The use of the cold-formed steel sheet bolted connections without washers is so significant; however, the North American Specifications for the Design of Cold Formed Steel Structural Members, NASPEC, doesn't provide provisions for such connections. The bearing failure of sheet and the shear failure of sheet were considered in this study. For the sheet shear strength, it was found that the NASPEC (2007) design provisions can be used for oversized holes in both single and double shear configurations and for the double shear connections on short slotted holes. For the sheet bearing strength, a new design method was proposed to be used for low and high ductile steel sheets. The method was compared with the NASPEC and the University of Waterloo approach. Washers were still required for single shear connections on short slotted holes. Besides, connections using ASTM A325 bolts yielded higher bearing strength than connections using ASTM A307 bolts.

In this research, monotonic and cyclic tests on cold-formed steel shear walls sheathed with steel sheets on one side were conducted to (1) verify the published nominal shear strength for 18-mil and 27-mil steel sheets; and (2) investigate the behavior of 6-ft. wide shear walls with multiple steel sheets. In objective 1: this research confirms the discrepancy existed in the published nominal strength of 27-mil sheets discovered by the previous project and verified the published nominal strength of 18 mil sheet for the wind design in AISI S213. The project also finds disagreement on the nominal strength of 18-mil sheets for seismic design, which is 29.0% higher than the published values. The research investigated 6-ft. wide shear wall with four framing and sheathing configurations. Configuration C, which used detailing, could provide the highest shear strength, compared to Configurations A and B. Meanwhile, the shear strength and stiffness of 2-ft. wide and 4-ft. wide wall can be improved by using the seismic detailing.

In cold-formed steel (CFS) construction, bolted connections without washers for either oversized or slotted holes may significantly expedite the installation process and lower the cost. However, the North American Specification (AISI S100, 2007) for the Design of Cold-Formed Steel Structural Members requires washers to be installed in bolted connections with oversized or slotted holes. A research project (Phase 1) sponsored by American Iron and Steel Institute (AISI) was recently completed at the University of North Texas (UNT) that investigated the performance and strength of bolted CFS connections with oversized and slotted holes without using washers. The research presented in this thesis is the Phase 2 project in which the bolted CFS connections were studied in a broader respect in terms of the failure mechanism, the material thickness, and the hole configurations. Single shear and double shear connections without washers using oversized holes, oversized combined with standard or slotted holes were experimentally examined. Combined with Phase 1 results, the Phase 2 gives a comprehensive evaluation of the behavior and strength of bolted CFS connections with oversized and slotted holes without using washers. Revisions to the existing AISI North American Specification requirements for bolted connections are proposed to account for the reduction in the connection strength caused by the oversized and slotted hole configurations without washers. Specific LRFD and LSD resistance factors and ASD safety factors for different hole configurations in terms of the new proposed methods were presented.

Cold-formed steel connections are commonly fastened using self-tapping self-drilling screws. The behavior of these connections can differ based on the screw manufacturer or the cold-formed steel product used, both of which have a large selection available for use in industry. Because of their popularity and the many possible variations of these connections, researchers have frequently tested screw connections to characterize their behavior. However, repeatedly conducting this type of experiment is time consuming and expensive. Therefore, the purpose of this work was to create finite element models that can successfully predict the behavior of single lap shear screw connections, a common connection type used in cold-formed steel framing. These models were created using the finite element program Abaqus/CAE. To validate these models, test results from Pham and Moen (2015) were used to compare the stiffness, strength, and failure mode of multiple connections. A parametric study is also conducted to determine the influence of contact parameters on the behavior of the model. The results showed that all models consistently had good agreement with the connection stiffness and that most of the models also had good agreement with the peak load and failure mode of the v tests. These results were also compared to the design equations available for screw connections from the American Iron and Steel Institute (AISI). This comparison revealed that the models are more successful at predicting screw connection behavior than AISI, and thus work is required to improve the accuracy of AISI’s design equations. The eventual goal of this work is to develop a procedure to build and validate models without requiring test data. This work continuing in the future can lead to recommendations to improve AISI’s design equations and to implement the behavior of the connections into large cold-formed steel framing models such as diaphragms or shear walls.

Thermal bridging can be defined as the phenomenon where a structural element spanning the building envelope acts like a thermal pathway which collects and moves energy (heat) from the interior to the exterior of the structure. CFS construction, due to the high thermal conductivity of steel with respect to its surrounding structural components and repetitive nature of framing, is highly prone to thermal bridging. Thermal bridging significantly alters the thermal performance of wall assemblies. Hence, the objective of this research project was to quantify the magnitude of energy loss through cold-formed steel (CFS) stud wall assemblies at a component level to lay the groundwork for future works that promote sustainable, energy-efficient, and improved building design recommendations. Therefore, a parametric evaluation was performed using ISO 10211:2007, Annex A, conforming heat transfer software Blocon Heat3 version 8 to generate the data required for analysis. 80 unique wall assemblies and the impact of selected parameters on the overall thermal transmittance of the wall assembly were studied as part of the parametric evaluation. The key variables of the study are steel thickness, stud depth, stud spacing, cavity insulation R-value, external insulation thickness (R-value), and fastener diameter and length. Based on the results of the analysis, effects of increasing stud and track thickness, depth, and stud spacing have been discussed in the form of trends in overall heat flow and linear thermal transmittance coefficient values. Additionaly, effects of increasing external insulation have been discussed by addressing changes in heat flow.

Cold-formed steel roof truss systems that use complex stiffener patterns in existing hat shape members for both top and bottom chord elements are a growing trend in the North American steel framing industry. When designing cold-formed steel sections, a structural engineer typically tries to improve the local buckling behavior of the cold-formed steel elements. The complex hat shape has proved to limit the negative influence of local buckling, however, distortional buckling can be the controlling mode of failure in the design of chord members with intermediate unbraced lengths. The chord member may be subjected to both bending and compression because of the continuity of the top and bottom chords. These members are not typically braced between panel points in a truss. Current 2001 North American Specifications (NAS 2001) do not provide an explicit check for distortional buckling. This dissertation focuses on the behavior of complex hat shape members commonly used for both the top and bottom chord elements of a cold-formed steel truss. The results of flexural tests of complex hat shape members are described. In addition, stub column tests of nested C-sections used as web members and full scale cold-formed steel roof truss tests are reported. Numerical analyses using finite strip and finite element procedures were developed for the complex hat shape chord member in bending to compare with experimental results. Both elastic buckling and inelastic postbuckling finite element analyses were performed. A parametric study was also conducted to investigate the factors that affect the ultimate strength behavior of a particular complex hat shape. The experimental results and numerical analyses confirmed that modifications to the 2001 North American Specification are necessary to better predict the flexural strength of complex hat shape members, especially those members subjected to distortional buckling. Either finite strip or finite element analysis can be used to better predict the flexural strength of complex hat shape members. Better understanding of the flexural behavior of these complex hat shapes is necessary to obtain efficient, safe design of a truss system. The results of these analyses will be presented in the dissertation. / Ph. D.

A general framework to design structural systems for a system-reliability goal is proposed. Component-based structural design proceeds on a member to member basis, insuring acceptable failure probabilities for every single structural member without explicitly assessing the overall system safety, whereas structural failure consequences are related to the whole system performance (the cost of a building or a bridge destroyed by an earthquake) rather than a single beam or column failure. Engineering intuition tells us that the system is safer than each individual component due to the likelihood of load redistribution and al- ternate load paths, however such conservatism cannot be guaranteed without an explicit system-level safety check. As a result, component-based structural designs can lead to both over-conservative components and a less-than-anticipated system reliability. System performance depends on component properties as well as the load-sharing network, which can possess a wide range of behaviors varying from a dense redundant system with scope for load redistribution after failure initiates, to a weakest-link type network that fails as soon as the first member exceeds its capacity. The load-sharing network is characterized by its overall system reliability and the system-reliability sensitivity, which quantifies the change in system safety due to component reliability modifications. A general algorithm is proposed to calculate modified component reliabilities using the sensitivity vector for the load-sharing network. The modifications represent an improvement on the structural properties of more critical components (more capacity, better ductility), and provide savings on less important members which do not play a significant role. The general methodology is applied to light steel-framed buildings under seismic loads. The building is modeled with non-linear spring elements representing its subsystems. The stochastic response of this model under seismic ground motions provides load-sharing, system reliability and sensitivity information, which are used to propose target diaphragm and shear wall reliability to meet a building reliability goal. Finally, diaphragm target reliability is used to propose modified component designs using stochastic simulations on geometric and materially non-linear finite-element models including every individual component. This material is based upon work supported by the National Science Foundation under Grant Nos. 1301001 (Virginia Tech), 1301033 (University of Massachusetts, Amherst) and 1300484 (Johns Hopkins University). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily re ect the views of the National Science Foundation. The author is grateful to the industry partner, the American Iron and Steel Institute, for their cooperation. / Ph. D. / This research proposes methods to design engineering networks for acceptable overall safety. Some examples of engineering networks include electrical systems, transportation systems and infrastructural systems. When any such system is designed, the properties of every individual component (size, capacity etc.) are assigned according to cost and safety requirements. However, it is typically very difficult to reliably quantify the overall safety of the entire system, which is technically known as ‘system reliability’. As a result, there are limited options for engineers to adjust the individual component designs within a system to achieve a pre-specified ‘targeted’ system reliability . This dissertation proposes computational and statistical methods to achieve this. The proposed methods are applied to a specific engineering system, namely a two story building subjected to ground shaking resulting from an earthquake. Computer models are developed for different scales of the building, beginning from the full building structure, then its individual floors and walls, and finally the individual components that make up each floor and wall. These models are verified with experimental results spanning all three scales. The verified models are then used to both compute the overall system reliability of the building subjected to earthquake ground shaking, as well as to modify its design component-by-component to achieve a targeted system reliability which is different from the system reliability of the original design. The results indicate that the as-designed reliability of the building system is adequate, but this reliability results from features of the building that are not expected to provide additional safety. The research demonstrates means to obtain this additional safety by redesigning the core functional building components, without relying on the unexpected added safety from ‘non-structural’ components (such as partition walls inside a building). The methods developed herein can be applied to redesign the components of various engineering system networks such that a targeted overall system reliability can be satisfied, resulting in improved performance and life-safety, potentially even at reduced costs.