Structures Congress 2006: 17th Analysis and Computation Specialty Conference

In traditional earthquake resistant design, the approach explicitly focuses on performance of the system for only a single design event level, which is generally termed "life safety". Contemporary efforts have attempted to provide more quantitative definitions of multiple performance levels in order to meet various expectations. Performance Based Seismic Design (PBSD) was proposed to provide reliable methods of meeting multiple performance goals. To progress toward the goals intended with PBSD, analytical and design techniques, evaluation procedures, and suitable test cases are needed to demonstrate and examine the reliability in a consistent manner. Thus, the ASCE Committee on Structural Control has recently developed a base isolated benchmark problem. The objective of this benchmark study is to provide a systematic and standardized means by which competing control strategies, including devices, algorithms, sensors, etc., can be evaluated in achieving the performance objectives. A framework for this study has been developed, from problem definition to sample controller design. Based on the framework built by the above researchers, this study aims to propose a approach and demonstrate a competitive design for a smart isolation system.Linear isolation bearings are the focus of the design procedure at this time. An H₂/LQG control algorithm is used to determine the optimal control action. MR dampers are adopted for the semiactive control devices, and the control input is determined by clipped-optimal algorithm. The behavior of this building is examined in terms of certain common evaluation criteria defined within the benchmark problem statement. The design is shown to be effective for reducing the responses.

This paper discusses the feasibility of using augmented situational visualization, commonly referred to as augmented reality (AR), to evaluate damage sustained by buildings in the aftermath of natural and human-perpetrated disasters. A study was carried out to evaluate the hypothesis that previously stored building information can be superimposed onto a real structure in AR, and that structural damage can be evaluated by measuring and interpreting key differences between a baseline image and the real view of the facility. Two experiments were performed in conjunction with structural tests that were being conducted to investigate the seismic performance of concrete walls.The obtained results highlight the potential of using AR for rapid damage detection and indicate that the accuracy of structural displacements measured using AR is a direct function of the accuracy with which augmented images can be registered with the real world.

A significant challenge for a damage identification (DI) method is the detection of a small fault. This is a key issue in determining what information sets should be utilized for damage identification. In order to locate damage, the feature for damage detection must contain local information of the system necessary for revealing localized anomalies caused by damage. Higher order derivatives of mode shapes are considered to be excellent features for damage identification due to their sensitivity to small faults. The use of second order derivatives of the mode shape variations of a structure to identify localized damage has been investigated extensively. A common limitation of all the DI methods used in the investigations discussed above is their dependency upon prior test data from the assumed healthy state of the structure for damage discrimination. Therefore, only structures with a known (or assumed) healthy state would be eligible for structural condition evaluation. This also complicates the analysis of subsequent test data, since variability in environmental and operational conditions can lead to yield false-positive and/or false-negative indications of damage. An issue of primary importance is the level of sensitivity of the second order derivatives to small levels of damage. Many researchers have validated their DI methods only on systems having relatively high levels of damage (e.g., stiffness losses of 10% or more). The main goal of a reliable DI technique must be the detection and location of damage at an early stage so that the structure can be repaired before damage grows to critical levels. In addition, a need remains to reduce the dependence upon engineering judgment to discriminate damage. This can be met through incorporation of statistical models to distinguish anomalies associated with damage to those as consequence of variability in testing procedures and errors in the data reduction process. The purpose of the HODD method is to provide a global experimental technique for nondestructive evaluation of beam-like structures that can reveal localized changes in structural properties without relying upon prior tests or complex numerical models of the structure. The method exploits the noticeable discontinuities that small changes in local stiffness introduce into mode shape derivatives. This leads to an efficient and robust identification of potential damage locations by looking for anomalous variations in a mode shape derivative pattern. Comparisons are made only within a given mode shape (and thus a given test), not across different tests, thus reducing the impact of environmental effects or changes in testing conditions. The HODD method is intended for use in conjunction with local non-destructive testing techniques that would assess more fully the extent of a fault located by this global technique, thus providing a detailed quantification of the severity of damage. The purpose of the present work is to present applications of the HODD method to experimental damage identification problems, illustrating some of the advantages and concerns regarding this approach. The method will be validated under a variety of damage scenarios, with consideration of highly localized damage cases. The case studies will demonstrate practical implementation of the HODD method.

Semiactive control has typically been designed for and applied to linear structures. Civil structures, however, are designed to yield, thus behaving nonlinearly during extreme dynamic loading. Because they cannot inject mechanical energy into the controlled system, semiactive devices are inherently stable and well suited for application to structures with aleatoric (the randomness of excitations) and epistemic (the inherent variability of structural and nonstructural components and systems) uncertainties and systems with the potential to behave nonlinearly. Additionally, the low power requirements of semiactive devices ensure that during extreme events, when external power may not be available, the semiactive device can continue to fully function using an alternate power source (e.g., a 180 kN Magneto-Rheological fluid damper can operate continuously in excess of 2 hours powered by an off-the-shelf universal power supply). Despite these advantages, semiactive control in the presence of nonlinear structural behavior has yet to be demonstrated experimentally. This paper will discuss the experimental setup for full-scale experimental verification of semiactive control strategies for buildings exhibiting nonlinear behavior during large seismic events utilizing the NEES shared-use Fast Hybrid Test system at the University of Colorado at Boulder. The experiments described in this paper will employ hybrid testing of three semiactive large-scale Magneto-Rheological (MR) fluid dampers while simulating in real-time the nonlinear response of a building structure subjected to suites of simulated and recorded earthquakes. This is the first application of fast hybrid testing to semiactive control devices. This paper first provides analytical results on the semiactive control of a nonlinear building model. Next, details to conduct Fast Hybrid Testing of semiactive control devices are presented and the experimental setup described. Lastly, conclusions and future work are identified.

Given the present burdens associated with inspection and maintenance of Civil Infrastructure, the development of effective, automated damage diagnosis techniques, including the sensor technologies that support them, has become a major research need. While recent developments in wireless sensor networks have demonstrated their potential to provide continuous structural response data to quantitatively assess structural health, many important issues including network lifetime and stability, damage detection reliability, and trade-offs in model order to balance computational capabilities must be realistically addressed. Only then can wireless embedded sensor networks become a practical tool for Structural Health Monitoring of large, complex Civil Structures. In response to these needs, the concept of a multi-scale wireless sensor network is introduced in this study with a restricted input network activation scheme and the integration of data from a heterogeneous sensor array to improve damage detection for low-order models. The multi-scale network concept introduced here helps to improve power efficiency, minimize packet loss and latency, and eliminate synchronization issues through the use of a decentralized analysis scheme and the activation of sub-networks only in the vicinity of suspected damage, while reducing the required size of the reference pool for the undamaged state. This study introduces the network architecture concept, a strain-driven approach to damage detection and preliminary simulated results.

Applying advanced hazard mitigation strategies such as structural control can provide protection against seismic excitations and allow engineers to build safer and more economical structures. Structural control increases the safety and performance of a traditional design by redistributing and dissipating the energy of the structure. Experimental verification of structural control can capitalize on the capabilities of NEES to conduct more extensive tests including, for example, the effects soil-structure interaction on the performance of structural control. Soil-structure interaction is the relationship between a structure and its supporting soil. The dynamics of the soil when combined with the dynamics of the structure can lead to significantly increased structural response depending on the frequency matching of soil and structure. One concern is the potential effect of this change in the system on the stability and performance of an applied control strategy. Full-scale or large scale investigation and validation of this complex, combined behavior is difficult to realize due to the expense and risks involved in the testing of such systems. Therefore, multi-site component testing, using interacting computational modules will aid civil engineering researchers in developing an understanding of such behaviors. Herein we demonstrate a step forward toward the realization of this goal. The focus is on an international experimental implementation consisting of a multi-site test performed using the NEES cyberinfrastructure and University Consortium on Instructional Shake Tables (UCIST) bench-scale shake tables controlled through the NEES network (NEESgrid). UCIST has been in existence since 1998 and was developed to facilitate educational efforts to demonstrate structural dynamics and earthquake engineering by introducing these topics at the undergraduate level. Originally consisting of over 23 universities, UCIST now encompasses over 80 shake tables around the world. Only recently was control of the UCIST shake tables through the NEESgrid accomplished at the Smart Structures Laboratory at the Colorado School of Mines. In the multi-site international experiment described in this paper, the effect of soil-structure interaction on the performance of a passive structural control strategy is examined. A substructure approach to soil-structure interaction is employed where the soil and the structure are tested in series. The shaking of the bedrock and alluvium is accomplished on one UCIST shake table at the Colorado School of Mines in Golden, Colorado. The controlled structure is tested on a second UCIST shake table at Nihon University in Tokyo, Japan. The multi-site experiment was successfully coordinated and controlled using the NEES cyberinfrastructure.

The hybrid simulation test method is a versatile technique for evaluating the seismic performance of structures by seamlessly integrating both physical and numerical simulations of substructures into a single test model. Using hybrid testing, complex structural systems composed of multiple large-scale experimental and numerical substructures can be evaluated by linking experimental facilities using the internet. In this paper, a distributed control strategy is presented that supports the implementation of hybrid testing with geographically distributed substructures using advanced algorithms that offer improved accuracy. The objectives are to provide a scalable framework for multiple-substructure testing at distributed sites and to improve the reliability of the test results by minimizing strain-rate and force relaxation errors in the remote experimental substructures. The control strategy is based on a multi-threaded simulation coordinator combined with an event-driven controller at the remote experimental sites. The effectiveness of this procedure is demonstrated by computing the earthquake response of a six-span bridge model with five remote experimental and numerical column substructures distributed within laboratories across the U.S. Further, the distributed tests were implemented using a secure network link between the testing sites that was developed for the NEES cyber-infrastructure.

This paper presents the development and implementation of a novel structural testing method involving the combined use of shake tables, actuators and computational engines for the seismic simulation of structures. The structure to be simulated is divided into one or more experimental and computational substructures. The interface forces between the experimental and computational substructures are imposed by actuators and resulting displacements and velocities are fed back to the computational engine. The earthquake ground motion is applied to the experimental substructures by shake tables. The unique aspect of the above hybrid system is force-based substructuring. Since the shake tables induce inertia forces in the experimental substructures, the actuators have to be operated in dynamic force control as well, since either the force or the displacement, but not both can be controlled at a given point and at a given instant of time. The substructuring strategy and the numerical integration algorithms associated with the computational substructures are presented along with the implementation of the computational engine. A new dynamic force control strategy developed for this purpose using series elasticity and displacement compensation is briefly reviewed. Issues related to time-delay compensation are also discussed. Finally, an example of a real-time hybrid test implementation, and results from this experiment are presented.

This paper presents the substructure testing methodology developed for a state-of-the-art fast hybrid test system and the testing of a steel zipper frame. The testing technique is based on the pseudodynamic test concept that combines model-based simulation with physical testing. In the hybrid tests presented here, only the bottom-story braces of a three-story zipper frame were tested, while the rest of the frame was modeled in a computer during a test using a general structural analysis framework OpenSEES. The tests have demonstrated the capability and reliability of the system. The discussion also covers pertinent issues and considerations for carrying out a successful test.

A system identification and model updating methodology that accounts for factors influencing the reliability of identification is proposed. An important aspect of this methodology is the generation of a population of candidate models. This paper presents an analysis of error sources that are used to define model populations. To interpret these populations we use data mining techniques such as correlation measurements and principal components analysis. These methods are useful for estimating the reliability of system identification when the number of candidate models is high.

This paper describes non-linear soil and structure analysis techniques fundamentals and application of this technique using the FBPier/FBMultiPier computer Program. A case study is discussed using the FBPier program on the 2,265 Meter long Benicia-Martinez Bridge located near The San Francisco Bay area in California. FBPier was used on this structure for the independent check during the design phase and through the construction to analyze changes in field conditions and foundation construction.

This paper investigates the consequence of using Rayleigh proportional damping in the analysis of inelastic structural systems. The discussion is presented theoretically, as well as by example through the analysis of a simple 5-story structure. It is shown that when the stiffness portion of the system damping matrix is based on the original system stiffness, artificial damping is generated when the structure yields. When the damping matrix is based on the tangent stiffness but the Rayleigh proportionality constants are based on the initial stiffness, a significant but reduced amplification of damping occurs. When the damping is based on the tangent stiffness and on updated frequencies based on this stiffness, virtually no artificial damping occurs. The paper also investigates the influence on effective damping when localized yielding occurs in areas of concentrated inelasticity. In such cases it is possible to develop artificial viscous damping forces that are extremely high, but that are not easy to detect. Such artificial damping forces may lead to completely invalid analysis. The paper ends with recommendations for performing analysis where the artificial damping is eliminated, or at least controlled.

Increased importance has been placed on the design and analysis of structures to account for potential terrorist attacks. Designers must account for many options that include various levels of mitigation and protection measures lo ensure the safety of the building and its occupants. One critical ingredient to the analysis is an understanding of the structural performance of structures subjected to blast loading especially when considering issues of collapse. The approach presented in this paper provides an efficient alternative to dynamic analysis of a moment-framed building subject to blast induced damage. First, framing members (girders and columns) are designed to withstand typical dead and live loads. Factored design loads are calculated using a first order (load path) static analysis of the frame. Structural steel, reinforced concrete, reinforced/unreinforced masonry, and wood members are designed in accordance to the applicable building code requirements and design specifications. After structural damage due to blast loads is predicted, progressive collapse of the building is modeled as a series of events. Events include local plastic hinge formation, member failure from overstress, local buckling. lateral torsion buckling, or axial buckling, and global buckling or instability. Element section yield and failure toads are evaluated for axial, strong and weak bending and torsion using the applicable codes (striped of load and resistance factors). Stiffness at hinges is assumed to be bilinear. Global buckling is evaluated by a generalized Eigenvalue approach and instability by simple Eigenvalues. The approach uses a course finite element model of the building that includes beams, columns, walls and floor slabs. The approach can determine if progressive collapse will propagate to the entire building or if a stable condition can be achieved. An optional model that includes the debris loading of removed structural elements is also included.

High-fidelity, explicit finite element analysis has proven to be an effective tool in simulating airblast effects on reinforced concrete structures. The challenge is to make intelligent predictions regarding the stability or failure of real RC structures using the results of these numerical simulations. Code guidelines, such as the Army's TM5-855, provide one aid in assessing RC member failure based on support rotation limits. Mechanics-based assessments using momentum and kinetic energy; fracture energy, volumetric, tensile, and plastic strains; and material damage levels provide additional information for predicting failure. Highly localized blast effects, such as cratering of the concrete, must also be taken into consideration. This paper will discuss such issues as it presents a general methodology for predicting states of damage and failure of RC structures due to blast loadings using high-end finite element analyses.

Under extreme temperature exposure thermal softening of stiffness and strength properties as well as thermal expansion and drying shrinkage of cement-based materials have come to the forefront in the safety assessment of tunnel and jet fuel fires in high rise buildings. It is indeed unfortunate that several accidents in the recent past have reinforced the need for a more comprehensive understanding of concrete and steel materials exposed to fire environments. Thereby the transient nature of rapid heating and cooling and the thermal mismatch between concrete and steel may introduce extensive spalling in reinforced concrete because of pore pressure build up in regions of severe temperature gradients. In this paper we explore the temperature sensitivity of concrete and steel and their effects on the thermal collapse of a reinforced concrete benchmark structure (RCS) representative of a fire scenario. Specifically we review the mismatch of thermal expansion and the thermal softening of the mechanical response behavior of concrete under heating and cooling. In the sequel, we illustrate some of these issues with the thermal collapse of a RC box structure subjected to an increasing temperature gradient. Whereas structural collapse is well understood using limit analysis concepts of plastic yielding in skeletal structures, failure under combined hygro-themio-mechanical load histories deserves more attention in view of the absence of upper and lower bound theorems for softening and transient conditions. Computational failure studies using finite element simulations introduce additional issues which give raise to fundamental questions what does constitute collapse, and how does numerical failure relate to physical collapse in the presence of viscous regularization and global stabilization strategies. In the last part of the paper we explore the classical collapse problem of a portal frame for which plastic limit analysis techniques provide well-established reference solutions to validate numerical failure simulations using the incremental damage-plasticity model for concrete in ABAQUS.

To maintain objective response without modifying the material stress-strain properties of a beam-column member, this paper presents a new element integration method that confines nonlinear constitutive behavior to plastic hinge regions of a specified length while maintaining numerical accuracy. The force-based formulation is ideal for this approach because the deformations in the plastic hinge regions are computed such that the corresponding section forces are in equilibrium with the element end forces. The paper begins with the force-based beam-column element formulation using the standard Gauss-Lobatto integration rule for distributed plasticity. Then, the concept of plastic hinge integration is presented, followed by Gauss-Radau quadrature, which forms the basis for the new element integration method. This paper concludes with a numerical example that shows the inclusion of a plastic hinge length in the element integration rule enables objective response for simulating beam-column behavior.

The objective of this paper is to promote the development of new hybrid optimization methods that combine simple algorithms based on structural engineering knowledge with sophisticated optimization algorithms. We illustrate the importance of this concept using algorithms that we have developed for different problems in structural engineering. First, we present a new algorithm for truss optimization that demonstrates a significant improvement compared to existing methods such as GAs. The performance of the algorithm is studied using widely popular examples in literature like the 10-bar truss. Then, we describe a GA-based approach with special adaptation for performing trade-off studies on discrete optimization problems with binary decision variables. The approach employs a unique crossover scheme that ensures that the offspring from the crossover have the same number of ones given that the two parents each had the same number of ones. The superior performance of this crossover over the traditional uniform crossover for trade-off studies is illustrated with application to certain examples. Finally, "MGA - Modeling to Generate Alternatives," (MGA) is used to generate alternatives that are closely-spaced in objective space but are farther-apart in decision space. MGA is different from GA in the sense that GAs evaluate alternatives only in objective space but not in decision space.

Typical models for isolation bearings use elastic-plastic (bilinear) or other empirically derived models for lateral force-deformation behavior. These models do not include the influence of axial loads on the lateral behavior, or more generally the interaction of lateral and vertical response as a result of geometric nonlinearities. Such effects have been shown to be well-represented by a combination of linear shear and rotational springs, i.e., the two-spring model. Here, the two-spring model is extended to consider material nonlin- earity in the shear spring, and an empirical representation of the experimentally observed variation of yield strength is included. The governing equations are reformulated to be compatible with a stiffness-based state determination procedure, in which the bearing forces are found by iterative solution of the nonlinear equilibrium and kinematic equa- tions using Newton's method, and the instantaneous or tangent bearing stiffness matrix is formed from the differentials of these equations. As an example, this model has been implemented as a material model for use with a zero-length spring element in OpenSees. Comparative response history analyses of slender isolated buildings demonstrate that the geometric nonlinearities have a significant influence on the peak axial forces in the the isolation bearings in strong ground motion.

This paper deals with the cyclic elastoplastic analysis and seismic performance evaluation of thin-walled steel tubular bridge piers. The basic characteristics of steel tubular bridge piers of thin-walled box and circular sections are noted. The results of a finite element analysis on cyclic elastoplastic behavior of steel bridge piers are presented. In the analysis the modified two-surface plasticity model is employed for material nonlinearity to trace with high accuracy the inelastic cyclic behavior of steel. Based on the ultimate strength and ductility capacity, a seismic performance evaluation method for thin-walled steel tubular bridge piers is presented. The procedure of determining ultimate strength and ductility capacity of piers is based on the empirical ductility equations for pier's plate sub-elements and involves an elastoplastic pushover analysis and failure criterion accounting for local buckling. The application of the method is demonstrated by comparing the computed strength and ductility of some cantilever columns with the test results. The method is applicable for both the design of new and retrofitting of existing thin-walled steel tubular structures.

Fiber Reinforced Polymer (FRP) decks have been successfully used to rehabilitate highway bridges in the U.S. One important characteristic of these decks is their inherent high strength to weight ratio. These decks also have low stiffness. Their design is thus governed by serviceability rather than strength. Meanwhile, preliminary analyses have suggested that existing serviceability criteria, intended to limit vibrations, are not applicable to FRP deck bridges. There are two primary reasons for this conclusion: first, the inherently low mass and stiffness of the FRP deck, and second, the deck-to-girder connections adopted in FRP deck bridge design are discrete and do not provide full composite action between deck and girder. Both properties will certainly produce a vibration response very different than that from a traditional reinforced concrete slab. In attempts to satisfy the existing serviceability criteria with FRP decks, more material and possibly a large number of deck-to-girder connectors are required. This results in expensive, impractical designs. Therefore, a joint analytical-experimental investigation is being conducted to establish appropriate serviceability criteria for FRP bridge deck applications. In this work, honeycomb, sandwich FRP decks are investigated. These honeycomb FRP decks have a very complex geometry, and computational limits prevent modeling of these bridge decks in detail. A simplified finite element modeling technique for this type of FRP deck was developed, as a first step, using the commercial finite element program ANSYS 8.1. In this paper, the modeling technique of each FRP bridge component is described in detail. This includes the use of the eccentric beam model for the steel girders and the steel guardrails. For validation of the proposed modeling technique, a detailed model of the honeycomb FRP deck was also created for comparison to the simplified model. Deflections, frequencies, and mode shapes for each model are compared with the experimental data. For validation of the overall bridge model concept, an existing bridge with concrete deck was modeled in a similar manner. The resulting frequencies and mode shapes are compared to those obtained in the field.

Performance based design can be defined as a design approach exclusively developed in order to have the bridge structure meet targeted performance goals for range of hazard levels. The design is primarily focused on meeting various performance objectives that are in line with the desired level of the service of the structure. Seismic performance criteria establish the safety and functional performance of the bridges after an earthquake. Movable bridges are an expensive key part of the transportation infrastructure and are usually irregular and asymmetrical in geometry and support conditions. Therefore it is important to have performance-based design as they present unique challenges and are required to be evaluated not only for stresses but also more importantly their performance. This should require the designers to deal with functionality of the complex connection of movable component and stringent mechanical tolerance that needs to be complied in the design of movable bridges to reduce losses after the seismic events.

Often, the design of a light-rail train bridge is governed by the imposed requirements of frequency, usually in the range of 2 to 3 sec^–1. This requirement produces large structural sections, much larger than required by strength or serviceability. The required range of frequencies is present to avoid uncomfortable riding for the train passengers when the train crosses bridges. In order to circumvent this stringent requirement, a particular dynamic analysis is required to show that vibrations are acceptable. However, the analysis is complex by the fact that train deformations depend on the deflected shape of the bridge and this, in turn, depends on the exerted force and displacements generated by the train. Thus, there is an interaction problem which is nonlinear even for linear geometry and material properties, due to the fact that the train-bridge system stiffness and mass vary with the position of the train (time). The coupling of the stiffness of the train and bridge is another source of nonlinerity. For design purposes, the use of specialized software which in most cases is train behavior-oriented becomes impractical and expensive. The idea of linearizing the problem using standard design software and moving the train (adding masses and springs) and recomputing train effects for each location becomes cumbersome and theoretically incorrect as is shown below. In design the dynamic effects are included by the impact factor of the order of 25%. Here we will show that for trains crossing at normal velocities that factor is adequate. This study uses a practical finite element approach. The effects of the interactions are introduced in the energetic formulation. The approach is general for any finite element used (beams, shells or solids), but here we restrain ourselves to the case of beam elements. The train is modeled as a system of masses, springs and dampeners lumped for the body and the wheels. The procedure described here can be easily generalized for other train configurations. The paper is organized as follows: Section 2 selects the degrees of freedom to be used and describes the variational (energetic) approach for the train-bridge system. Section 3 explains the numerical formulation and addresses the importance of damping in the response. Section 4 presents the application of the methodology to a real design. Finally, Section 5 presents the conclusion and future research.

The trend in the architecture/engineering/construction industry toward the use of three- dimensional representations of structures in design, analysis and construction has led to the acceptance of the Computer Integrated Manufacturing (CIM) Steel Integration Standards (CIS/2) format for steel structures. This allows all information defining the structural steel to be passed among the structural designer, detailer, fabricator, and erector, as well as the other members of the project team, through a digital data file that eliminates the need for reentering information thus minimizing the potential for errors in translation. The three-dimensional model also facilitates the task of checking for interferences and inconsistencies which will reduce these problems during construction. This technology will continue to gain acceptance and must be developed for all components of the constructed project. A format is proposed for describing the geometry of typical reinforced concrete structures as a function of user-defined parameters. The structure is constructed from solid Comer, Beam and Slab components aligned with a flexible grid. Comers are analogous to finite element nodes while Beams and Slabs would be one- and two- dimensional elements, respectively, in conventional structural models. Once the number of grid spaces is defined in the X, Y and Z directions, the template designates which grid "squares" are filled with a Slab and which are open. Slabs may be floors or walls and may have holes. Beam components are aligned with grid lines and may represent beams, columns, or elements that connect adjacent Slabs. Comers are hexahedrons at the intersection of grid lines. The dimensions of the components are defined as functions of parameters listed in the template. A simple application of this approach would be a template to model a multiple story concrete frame structure. Advanced commercial structural analysis programs such as RAMSteel® and ETABS® allow parametric input to quickly generate this type of structural model. A template for this structure could be designed with minimal parameters if column spacing and floor spacing were constants and all structural members had the same dimensions. This basic template can then be enhanced to give the designer more flexibility throughout the structure by adding additional parameters. The result is a solid model which can be used as a basis for a finite element analysis and structural design but also for detailing rebar, checking interferences, designing formwork, planning construction, exporting to another 3D design environment for architectural and mechanical/electrical/plumbing design, and for generating two-dimensional drawings of the structure. In the future solid models will be part of the permanent record of the constructed project. Two applications of the template format are developed: a simple reinforced concrete frame and a more complex reinforced concrete pumping station. The graphical user interface allows the designer to change any dimensional parameter to immediately update the structure geometry in order to meet the project requirements. The final solid model is then converted to a finite element model and analyzed to determine shear, moment and axial force in beams, columns and slabs. Algorithms are being developed to use these results to design the reinforced concrete members. A sampling of the results of the analyses is presented.

In the present study, simple empirical formulae have been proposed for the reduction in bond strength as a function of rebar corrosion. These formulae have been proposed by considering a wide range of the experimental data pertaining to pullout testing and flexural testing of RC members. The formulae are then evaluated through their ability to reproduce the experimental data. A methodology has also been proposed to evaluate the flexural strength of the corrosion damaged RC beams failing in bond. The predicted flexural strengths from the proposed study are then compared with the available experimentally observed and analytically predicted flexural strengths of other researcher. The comparison shows that the proposed models are practicable in predicting the residual bond strength of corroded reinforcements.

Concrete wall panels have traditionally been cast using a rigid formwork. Straightforward methods of analysis and design are available for the traditionally cast concrete wall or floor panel. This is not so for the panel cast in a flexible fabric formwork. To date, no design procedures or methods to predict the deflected shape of a fabric cast panel have been developed. This paper introduces a design procedure that allows one to design a fabric cast concrete panel. A four-step procedure for analytically modeling a fabric formwork was developed employing the structural analysis program ADINA to analyze the formwork and the concrete panel cast in it. The final panel form, function and performance of the fabric membrane and the reinforcement of the panel for design loads all add to the complexities of the panel's analysis and design. Analytical modeling and design techniques presented in this paper will allow members of the design community another way to express themselves using a flexible fabric formwork. No longer will designers feel constrained by the limitations imposed by using a rigid formwork.

The most efficient method of replacement of major internal components such as steam generators in nuclear power plants may require the creation of a construction opening in the side of the containment building. For repairing post-tensioned concrete containment buildings, a major design challenge is to develop the most efficient scheme for tendon de-tensioning and subsequent re-tensioning. Stresses and displacements must be monitored in the containment wall and liner throughout the repair sequence. Nonlinear finite element analysis using the ANSYS general analysis package can be used to simulate the entire construction process and thus assure an adequate design margin of safety at all stages. Recent developments in computer CPU speed and RAM advancements have made it possible to perform complex nonlinear Finite Element Analysis (FEA) on an entire containment building overnight on a desktop machine. The nonlinear analysis technique discussed in this paper includes explicit modeling of the tendons and concrete including the tendon-concrete load interaction. Tendon tensioning and de-tensioning is modeled using an initial strain approach where link elements are coupled to the containment building wall. The wall is modeled with 3-D brick elements. Element birth and death is used to simulate the process of cutting the construction opening and subsequent patching of the wall in a stress-free state. This modeling method is critical to capture the local bending response in the patch that is often neglected in simplified models. This paper presents the details while illustrating the importance of explicitly modeling the construction and hole-patch-wall interaction.

The dynamic stiffness matrix for the spatially coupled free vibration analysis of thin-walled curved beam with non-symmetric cross section on two-types of elastic foundation is newly presented based on the power series method using the technical computing program Mathematica. For this, the elastic strain energy considering the axial/flexural/torsional coupled terms and the kinetic energy including the rotary inertia effect and the energy due to the elastic foundation are introduced. Then, equations of motion and force-displacement relations are derived from the energy principle and explicit expressions for displacement parameters are derived based on power series expansions of displacement components. Finally, the exact dynamic stiffness matrix is determined using force-displacement relations. In order to demonstrate the validity and the accuracy of this study, the natural frequencies of thin-walled curved beams with mono-symmetric and non-symmetric cross sections are evaluated and compared with the analytical solutions and finite element solutions using Hermitian curved beam elements and ABAQUS's shell elements. In addition, some results by the parametric study are reported.

The paper aims to study the flexural response of the foundation of a surface reservoir resting on a dense granular foundation bed overlying a weak clay layer with the reinforcement possessing sufficient bending stiffness placed at their interface. The footing and reinforcement are idealized as free-ended beams. The overlying compacted granular layer and the underlying clay layer are idealized respectively by Winkler Springs and four-element standard visco-elastic Burger model. A quadratic variation of the subgrade modulus is assumed along the length of the footing at the soil-structure interface with the corresponding maximum value being at the centre and progressively decreasing to minimum at the edges. The resulting differential equations are discretized by using Finite Difference Method (FDM), and the set of linear equations so obtained is solved by Gauss-Seidel iterative technique. Studies were carried out to obtain the optimal element size for convergent solution and establish its correctness by comparing the same with available solutions. Parametric studies to find the effect of time, relative stiffness and viscosity, interface friction and the distribution of subgrade reaction on the flexural behavior of the footing are also reported here.

In this paper, an attempt is made to develop a bridge management system that can provide practical maintenance plans by using a multiple objective genetic algorithm. In order to find out several useful solutions from a set of Pareto solutions, a 3D graphical system is developed by using JAVA techniques.

In order to establish a rational bridge management program, it is necessary to develop a cost-effective decision-support system for the maintenance of bridges. In this paper, an attempt is made to develop a bridge management system that can provide practical maintenance plans by using an improved multi-objective genetic algorithm. A group of bridges is analyzed to demonstrate the applicability and efficiency of the proposed method.

Deterioration mechanisms of existing structures have been noticed and studied over decades. Uncertainties associated with mechanical loadings and environmental stressors make it difficult to predict the life-cycle performance of these structures. In general, deteriorating structures are maintained by periodical interventions based on their condition states. Available resources are limited and maintenance decisions resulting from management systems concentrating on condition states of deteriorating structures are not always cost-effective. Therefore, the need for reliability-based structure management is evident. Models for time-based and performance-based (condition-based or reliability-based) maintenance strategies are applied in this study to a deteriorating bridge in Colorado. Several maintenance interventions are considered and combined. Realistic data consisting of condition, reliability, and cost of maintenance actions are used. Finally, optimization to minimize the total expected maintenance cost under multiple maintenance actions is performed. The optimum maintenance strategy considering condition, reliability and cost is selected based on different criteria.

Two competing methods were introduced for analyzing the condition of a structural system. The results produced by both were very different. In the weighted average approach, the system condition index will always be somewhere between the condition of the best and worst component in the system. In the traditional reliability approach, the condition index will always be lower than the condition of the worst member in a series system and higher than the condition of the best member in a parallel system. A traditional reliability approach works extremely well for a strength-based system where the importance factors of the components are relatively equal and the consequences of failure are typically catastrophic. In a serviceability context where some failures are more serious than others and some components are clearly more important than others, the information provided is less useful. The extreme values obtained in the traditional approach exaggerate the condition of the structure. A condition index of over 100 for the parallel structure would probably ensure it does not get replaced until every portion of the system is significantly deteriorated. In a multi-tiered series system, such as the spillway gate system presented here, the condition index would be so low that it would appear that every structure was in dire need of replacement. The condition of the worst element in the system, no matter how minor, controls the maintenance decision. If the goal is to use the overall condition of a structure to prioritize and optimize maintenance funding, the weighted average approach seems to provide better decision making information. By combining the condition of components with their importance to the overall system, it is easier to make a distinction between competing priorities. While series and parallel systems are treated in the same manner, a distinction could be made in the assignment of importance factors where a redundant system might receive a lower importance factor.

This paper outlines and demonstrates structural reliability and other gradient-based applications available for use in OpenSees. The Open System for Earthquake Engineering Simulation is an open-source, object-oriented finite element software framework developed for performance-based earthquake engineering analysis. OpenSees began as the computational platform for seismic simulations of structural and geotechnical systems in the Pacific Earthquake Engineering Research Center (PEER). Parallel computing, database, and hybrid simulation capabilities are included in the OpenSees framework making it an ideal environment for network-based simulations, e.g., in the NSF-sponsored George E. Brown Jr. Network for Earthquake Engineering Simulation (NEES). From this introduction of the framework in OpenSees for gradient-based applications in performance-based engineering, the objective of this paper is to summarize the top-level reliability and sensitivity computations in OpenSees. A listing of specific finite element modules available for gradient computations in OpenSees is provided, followed by representative examples of response sensitivity analysis and probabilistic reliability assessment.

Using reinforced concrete retaining wall as an example, the purpose of this paper is to demonstrate that one does not have to be an optimization expert to take advantage of the power built into modern tools such as the Microsoft Excel. Worksheets designed for usual analysis and design computations can be turned into formal design optimization tools by simply choosing cells corresponding to design variables, objective function and suitable constraint functions. This way an optimum design can be obtained in a systematic manner without having to resort to manual trial and error process. The paper presents the worksheets for design of cantilever retaining walls and shows several optimum design examples. To further demonstrate the usefulness of these worksheets a sensitivity analysis is carried out that shows the effect on optimum cost of various parameters such as wall height, surcharge pressure, allowable soil pressure, and internal soil friction angle.

Microsoft Excel offers users the ability to create custom functions that can be used in spreadsheets in the same manner as built-in functions. The key to this process is the use of a programming language included in Excel called Visual Basic for Applications (VBA). This paper discusses methods for creating useful custom functions, strategies for getting the most out of the functions, and examples of custom functions for structural steel design.

Most structural engineers use Microsoft (MS) Excel in the office on a daily basis. Using a 90-story reinforced concrete building currently under construction as an example, this presentation will illustrate some more unique ways in which Excel can be used as a very powerful tool in structural design. The tools shown are focused around some simple code written in Visual Basic for Applications (VBA). Please note that the author is a structural engineer, not a programmer, and has no formal VBA training. The code is actually quite simple for any engineer to pick up, especially since Excel will "record" most commands for the user to give a flavor for the syntax of the code. This presentation can be divided into four main sections: (1) an engineer's introduction to VBA, (2) custom functions in concrete design, (3) sample "tools" created in Excel and applied to a 90-story reinforced concrete building, and (4) appendices containing sample code and other useful information about VBA.

In the design of a long span roof system comprised of tubular steel members it was found that the structural member design is governed by connection performance. Existing codified guidance on the design of these connections was found to be unsuitable and an alternative approach was adopted. This approach follows a capacity/demand methodology where demands are determined using a global beam model of the entire structure and 2^nd order effects are included as determined by using an eigenvalue approach to amplify design forces. Connection capacities are established from moment- rotation curves developed by carrying out virtual tests on connections using non-linear finite element analyses.

A basement in an office building has been used for a storage or an instrument spaces. In addition, in a residential house it will increasingly utilize as a space for the particular purpose considering a blocking of a noise and light. In such case, a steel retaining wall is one of the choices to construct the retaining system because of its light-weight and its easy fabrication comparing with R/C retaining wall. The steel retaining wall system is subjected to an axial compression by self-weight of the building and to lateral bending and in-plane bending as well as shear by the earth and hydraulic pressure. In the previous paper the stiffened panel was analyzed under the compression and in-plane shear numerically and experimentally to apply the structure to the retaining wall for the housing basement. However, the steel retaining wall should be designed under such complicated combination of the loading mentioned above. In addition, the steel retaining wall must be fabricated with a stiffening system to obtain the sufficient strength of them. In this paper, the stiffened panel is analyzed under the combined compression and in-plane shear or the combined compression and bending numerically. In numerical analysis, the finite element methods are employed. The steel panel element, such as plate panel and stiffeners, are divided into degenerate shell elements and each element is subdivided into layers to represent the elasto-plastic behavior of each elements. In the numerical calculation, the lateral pressure is introduced under displacement increment scheme after a self-weight is introduced gradually. And several stiffened panels are analyzed to represent the applicability of the retaining wall system to the housing. From the numerical analyses, these behaviors show sufficient stiffness under a combined compression and in-plane loading and also the panels under a combined compression and bending show smaller ultimate strength than those of the combined compression and in-plane loadings. It is concluded that the appropriate combination of stiffened panels brings us a sufficient retaining wall system. Finally, combining the sample experimental analyses, it is concluded that the appropriate combination of stiffened panels brings us a sufficient retaining wall system.

Three methods exist to represent the reinforcement bars namely discrete representation, layered (or distributed) representation, and embedded representation. The main aim of this paper is to develop a simulation for the reinforcing steel bars in reinforced concrete using three-dimensional finite elements. The weighted residual method is used to formulate the finite element equations. Eight nodes isoparametric brick element (linear brick elements or Q8) are used to represent the concrete structures. The contribution of steel reinforcement to the stiffness matrix is achieved directly by using bar stiffness matrix, without undertaking numerical integrations that used in other methods. Since isoparametric brick elements are used in this model, the nodes of the truss element of the steel bar can be located using the global coordinates of the brick element nodes. The nodes of the truss element that represent the steel bars can be located in any location within or on the boundary of the brick element. According to that, embedded truss elements that represent the steel bars can be located in any location in the brick element, can be in any size and can be in any orientation. Moreover, the proposed method overcomes the difficulties of the numerical integrations and the computer programming.

An important problem is the development of a thermo-mechanical model of plasterboard- lined partition submitted to fire loads. A fundamental key to solve such a problem is the development and the experimental validation of a deterministic and a probabilistic thermo-mechanical model of cardboard-plaster-cardboard (CPC) submitted to fire loads. The proposed model takes into account data and modeling uncertainties. This research is justified by the fact that fire resistance tests of plasterboard-lined partition are made impossible when their dimensions exceed those of furnaces. This research is organized in four steps. The first one concerns the constitution of an experimental thermo-mechanical data base for a CPC multilayer and for its components. These experimental tests are carried out by the use of a bench test specially designed for this work. This device reproduces a heat flux equivalent to the one produced during a mandatory test using a gas furnace. The second step is the development of an homogenization thermo-mechanical mean model for the CPC multilayer. This mean model is adapted to a temperature range in which plaster or cardboard may be damaged. The third step is devoted to the implementation of the probabilistic model to take into account data and model uncertainties, in the last step, we present the results of numerical simulations that are compared to the experimental data.

The paper describes issues involved in a novel computational approach to non-linear structural analysis. The method is called a physics engine, which has been widely used in computer games and other visually-oriented simulations, but little used in structural engineering. The paper describes an implementation which models a structure as a collection of point masses in a time-step simulation. The position and velocity of each particle are known at the beginning of the time step, and the method calculates forces on the particles resulting from element deformation, gravity, damping, and other sources. The forces for each degree of freedom are then divided by the corresponding mass quantity to calculate acceleration. The acceleration is used in solving the differential equations of motion to determine the position and velocity at the end of the time step, and the process repeats. This procedure has the following distinguishing characteristics: all calculations are done with respect to the deformed structural geometry; there is no global stiffness matrix; mass must be assigned to all degrees of freedom; the method is extremely computationally intensive. The large displacement characteristic means that the method is well suited to modeling phenomena such as post-buckling behavior. The lack of a global stiffness matrix means that the method can easily analyze unstable structures and mechanisms. The combination of these characteristics makes the method well suited to modeling collapse-related phenomena resulting from removing members. The requirement to assign mass to all degrees of freedom raises questions about the proper calculation of rotational mass, which the paper discusses. Because the method is computationally intensive, it is limited to relatively small scale problems. It has been implemented in a program called Arcade, and has been primarily used for teaching, since small-scale models are effective for teaching. The paper briefly presents some teaching applications.

DLB can highly improve the parallel solution of nonlinear finite element problems. However, to integrate DLB algorithms into finite element problems, the solution method must be amenable to the DLB integration. In this work, a hybrid solution approach to parallel nonlinear finite element is introduced and evaluated. Results show that the proposed solution algorithm is highly efficient, with superlinear speedups observed in most test cases. However, solution efficiency of a given problem varies according to three main factors: problem complexity, number of processors, and communication cost. Each of these factors is dictated by numerous variables. However, acknowledging these factors paves the way for developing an integrated DLB that minimizes their effects in order to attain even higher speedups and efficiencies. Such system can highly improve the performance of current finite element solvers for nonlinear problems.

Throughout the past four decades, researchers (including this author) have developed a variety of structural optimization algorithms that produce a single optimum design with each execution. These algorithms range from optimality criteria methods to gradient-based nonlinear programming to stochastic methods. Most of these algorithms begin with a starting design and iterate to produce a final design. When a designer receives a single optimum design generated from an optimization algorithm, he/she can either: 1) accept the design and use it, or 2) reject the design and seek a design by means other than optimization. It has been the author's experience in working with designers that the second option is taken more often than the author would prefer. One reason designers reject the algorithm-produced optimum design is a psychological reason: designers are simply uncomfortable using a design produced by a computer algorithm. They feel that the final choice of design shouldn't be made by a computer. A second reason is that when the designer inspects the algorithm-produced optimum design, it is often readily apparent that important design criteria were not considered in the mathematical formulation of the optimization problem; criteria such as simplicity, aesthetic beauty, redundancy, constructability, etc. Such criteria are often subjective and difficult to formulate mathematically. Imagine an optimization algorithm that provided the designer with a choice of designs. Of course, one would expect these designs to be "optimal" or "near optimal" in some sense. One would also hope that these designs showed significant variation one from another. Such a variety of good designs could inspire ideas within the designer. The range of designs would also give the designer a better understanding of what is realistically possible. The designer's ultimate selection could include subjective criteria. Furthermore, the human designer would be psychologically comfortable in his/her role as decision-maker as he/she selects, combines, or modifies the designs produced by the computer. The computer assumes its proper role as preprocessor to human decision-making. The author believes that optimization algorithms that give designers a choice would make optimization more attractive to industry.

The optimization of large trusses often leads to a nearly optimal solution, rather than a truly optimal design. In fact, the problem space for truss optimization grows exponentially with the size of the truss. Using the method of problem reduction, this paper demonstrates that truss optimization is in the set of NP-complete problems. Hence, the only practical techniques for solving the truss problem are heuristic in nature. Genetic algorithms provide a viable solution for large trusses.