«R. SHANKAR NAIR R. Shankar Nair R. Shankar Nair, Ph.D., P.E., S.E. is a principal and senior vice president of Teng & Associates, Inc. in Chicago. In ...»
∆ Figure 3. Performance assessment procedure of FEMA 273/274/356 (after Comartin) Other projects including ATC-40, Methodology for Evaluation and Upgrade of Concrete Buildings and Vision-2000 Framework for Performance-based Seismic Design Project further developed and extended the technology developed in FEMA-273/274. These technologies were further refined by the American Society of Civil Engineers in their conversion of the FEMA-273/274 reports into the Prestandard for Seismic Rehabilitation of Buildings, FEMA-356 (FEMA 2000a). Together, the FEMA-356, ATC-40, and Vision-2000 documents define the current state of practice of performance-based seismic engineering.
Hamburger (2003) identified key shortcomings with the state of practice characterized by FEMA 273/356, including (1) the current procedures predict structural response and demands based on the global behavior of the structure but evaluate performance on the basis of damage sustained by individual components with the result that the poorest performing elements tend to control the prediction of structural performance, (2) much of the acceptance criteria contained in the documents, and used by engineers to evaluate the acceptability of a structure’s performance is based on judgment, rather than laboratory data or other direct substantiating evidence, leading to questions regarding the reliability of the procedures, (3) many structural engineers view the guidelines as excessively conservative, when compared against designs developed using prescriptive criteria, however, the reliability of the guidelines and their ability to actually achieve the desired performance has never been established, and (4) the performance levels of FEMA 273/356 do not directly address some primary stakeholder concerns, that is probable repair costs and time of occupancy loss in the building, due to earthquake induced damage.
Following the discovery of unanticipated brittle fracture damage to welded moment-resisting steel frame buildings following the 1994 Northridge earthquake, FEMA sponsored a large project (widely known as the SAC Steel Project) to develop seismic evaluation and design criteria for that class of buildings. Key products of the project included a series of recommended design criteria documents [FEMA-350 (FEMA 2000c), FEMA-351 and FEMAwhich incorporated performance-based design methodologies that addressed some of the issues associated with the state of practice per FEMA 273/356. The FEMA technical reports provided a large database of research data on the structural performance of this one structural system, which permitted the development of less subjective acceptance criteria for use in design, developed a methodology for evaluating the structural performance of a building based on its global response and behavior rather than solely on the amount of damage sustained by individual structural components, and developed a methodology for characterizing a level of confidence associated with a design’s ability to meet a performance objective, addressing in part, concerns related to designer warranties of building performance (Hamburger 2003). Although the SAC performance methodology has not seen widespread acceptance, the prescriptive procedures contained in FEMA-350 and FEMA-351 that were validated using the performance-based methodology have been widely accepted and incorporated into national design standards and building codes.
Towards performance-based earthquake engineering FEMA has contracted with the Applied Technology Council (ATC) to develop a next generation of performancebased seismic design guidelines for buildings, a project known as ATC-58. The guidelines are to be applicable to new and retrofit building construction and will address structural and non-structural components. Although focused primarily on design to resist earthquake effects, the next generation performance guidelines will be compatible with performance-based procedures being developed at this time for other hazards including fire and blast.
The ATC-58 project will utilize performance objectives that are both predictable (for design professionals) and meaningful and useful for decision makers. Preliminary project work tasks have revealed that these decision makers (or stakeholders) are a disparate group, representing many constituencies and levels of sophistication (Hamburger 2003). Decision makers include building developers, corporate facilities managers, corporate risk managers, institutional managers, lenders, insurers, public agencies and regulators. Each type of decision maker views performance from a different perspective and select performance goals using different decision making processes.
The performance-based design methodology will include procedures for estimating risk on a design-specific basis, where risk will be expressed on either a deterministic (scenario basis or event) or a probabilistic basis. Risk will be expressed in terms of specific losses (e.g., cost of restoration of a facility to service once it is damaged, deaths and downtime) rather than through the use of traditional metrics (e.g., life safety in a design-basis earthquake).
The performance prediction process is similar to that utilized in the HAZUS national loss estimation software, although the individual steps in the process will be implemented differently. Figure 4 from Hamburger (2003) is the flow chart for the ATC-58 performance prediction methodology. Much of the methodology is based on procedures currently under development by the Pacific Earthquake Engineering Research (PEER) Center (Moehle 2003) with funding from the U.S. National Science Foundation.
Figure 4. ATC-58 performance prediction flowchart (Hamburger 2003)
The PEER performance-based methodology is formalized on a probabilistic basis and is composed of four sequential steps: hazard assessment, structural/nonstructural component analysis, damage evaluation, and loss
analysis or risk assessment. The product from each of these four steps is characterized by a generalized variable:
Intensity Measure (IM), Engineering Demand Parameter (EDP), Damage Measure (DM), and Decision Variable (DV), for each of the steps, respectively. Figure 5 illustrates the methodology and its probabilistic underpinnings.
The variables are expressed in terms of conditional probabilities of exceedance (e.g., p( EDP | IM ) ) and the approach of Figure 5 assumes that the conditional probabilities between the parameters are independent. Moehle (2003) and Hamburger (2003) describes the performance-based methodology that has been adopted for the ATC-58 project. Key features of the methodology as presented by Moehle and Hamburger are summarized below for a building of a given geometry and design (termed D in the figure) and location (termed O in the figure). As such, the building and the hazard are fully defined.
Figure 5. Probabilistic underpinnings of the PEER and ATC-58 performance methodologies (Moehle 2003) Hazard analysis is the first of the four steps and produces ground motion Intensity Measures (IMs).
The traditional IMs are peak ground acceleration and spectral acceleration at selected periods. Values for the IMs are obtained by probabilistic seismic hazard assessment at the location of the site (O). IMs are typically described as a mean annual probability of exceedance of the IM ( p IM in Figure 5). The second step in the process is to use the IMs and the corresponding earthquake histories for simulation of the building response and the estimation of Engineering Demand Parameters (EDPs). EDPs, which traditionally have included component forces and deformations and story drifts, are calculated by linear or nonlinear methods of analysis. The products of the analysis are a conditional probability, p( EDP | IM ), for each EDP, which are integrated over the p IM to estimate the mean annual probability of exceedance of each EDP, p EDP. The third step in the process is to relate the EDPs to Damage Measures (DMs) that describe the physical state of the building. The outcome of this step are conditional probabilities, p( DM | EDP), which can be integrated over the p EDP to calculate the mean annual probability of exceedance for the DM, p DM. The fourth and final step in the PEER_ATC-58 methodology is to calculate Decision Variables (DVs). The mean annual probability of exceeding a DV, p DV, is calculated by integrating the conditional probability p( DV | DM ) (or loss function) over the p DM. The PEER_ATC-58 methodology can be expressed in terms of a triple integral of (2) based on the total probability theorem, namely,
where all terms have been defined previously. Column 2 of Table 2 below lists IMs, EDPs, DMs and DVs that could be adopted by the ATC-58 project for steel moment-frame construction. Column 3 lists similar measures that could be applied in the case of blast engineering.
Introduction Prior to the mid-1990s, analysis and design of building structures in the United States to resist the effects of blast loading and progressive collapse was undertaken by a relatively small group of specialty design professional consultants for a limited number of clients that managed high-exposure facilities such as government buildings, courthouses, and defense- and energy-related structures. Mainstream structural-engineering consultancies were not involved in blast engineering. The terrorist attacks on the World Trade Center in 1993 and 2001, the Murrah Building in 1995, and the Pentagon in 2001 altered substantially the attitude of the structural engineering community, building owners and insurers regarding blast design of commercial building construction, and there is renewed design-professional interest in blast engineering. However, because the blast-engineering designprofessional community is smaller than the earthquake community and blast considerations in commercial building design were the exception rather than the norm, there has been no national effort, on the scale of the FEMA-BSSC effort for earthquake engineering (FEMA 2000b), to produce guidelines and commentary for the analysis and design of blast- and progressive-collapse-resistant buildings.
The General Services Administration (GSA) has developed guidelines for progressive collapse analysis and design for new federal office buildings and major modernization projects (GSA 2003) but these guidelines are for limited distribution at the time of this writing. The GSA guidelines represent the state-of-the-practice in blast engineering of buildings but, similar to current building codes for seismic design, make use of indirect methods of analysis and prescriptive procedures of unknown reliability (Hamburger and Whittaker 2003). Resource documents for blast engineering are being developed currently by FEMA (FEMA 2004a, 2004b) but these documents will not provide explicit guidelines for analysis and design.
Performance-based blast engineering should make possible a process that permits design and construction of buildings with a realistic and reliable understanding of the risk of loss (physical, direct economic and indirect economic) that might occur as a result of future attack. This process could be used to (a) predict the global response, degree of damage (and perhaps economic loss) to a building subjected to a scenario blast event (Figure 6a) or physical attack (Figure 6b); (b) design individual buildings that are more loss-resistant than typical buildings designed using prescriptive criteria of a documents similar to GSA (2003); (c) design individual buildings with a higher confidence that they will actually be able to perform as intended for a blast attack; (d) design individual buildings that are capable of meeting the performance intent of the prescriptive criteria, but at lower construction cost than would be possible using the prescriptive criteria; (e) design individual buildings that are capable of meeting the performance intent of the prescriptive criteria, but which do not comply with all of the limitations of the prescriptive criteria with regard to configuration, materials and systems; (f) investigate the performance of typical buildings designed using prescriptive provisions and develop judgments as to the adequacy of this performance; and (g) formulate improvements to the prescriptive provisions so that more consistent and reliable performance is attained by buildings designed using prescriptive provisions.
Towards performance-based blast engineering
Components of equation (2) are broadly applicable to performance-based engineering for all loading conditions, normal and extreme. Significant overlaps should exist for extreme blast and earthquake loadings because inelastic response of the framing system is anticipated in both cases. That said, there are significant differences between blast and earthquake engineering in the loading environment (hazard or IMs) and important differences in simulation procedures and component response (EDP) assessment.
Blast loads on building structures (Biggs 1964; Mays and Smith 1995; Conrath et al. 1999) produce fundamentally different component responses than earthquake shaking. Further, blast loads are characterized deterministically at this time using scenario events ( α charge weight at β distance) and not probabilistically using a hazard curve as described in the previous section. Using the terminology associated with equation (2), sample IMs for blast loading are listed in Table 2 above for explosives placed outside and inside a building. Similar to the translation of earthquake IMs into earthquake histories for the purpose of simulation, blast IMs must be transformed into loading functions, including pressure-impulse curves for assessing component integrity (response) to direct air-blast and the likelihood of component loss; and loading functions associated with component elimination due to air blast.
Simulation of building response to earthquake shaking involves subjecting a (nonlinear) mathematical model of the building frame to one or more earthquake histories. Nonlinear component models for such simulation should be based on experimental test data similar to that collected systematically in the SAC Steel Project (FEMA 2000c).
Two assumptions made in the SAC Steel testing program and for the development of beam-component models were that beams are deformed primarily about their strong axis and twisting and distortion of the gross section is avoided.