Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 28
28 and need to be generated or selected (Payer et al., 1994). Tar- heavy timber, and masonry structures, as well as discussions get reliability levels vary from one application to another due of issues that should be considered when making the cali- to various factors, including implied reliability levels in cur- brations. Table 13 provides typical values for T based on rent design practice, failure consequences, public and media values provided by Ellingwood et al. (1980). The target reli- sensitivity, types of users and owners, design life of a struc- ability levels shown in Table 14 are used by Ellingwood and ture, and other political, economic, and societal factors. For a Galambos (1982) to demonstrate the development of partial general view, see Whitman (1984) and Becker (1996). Two safety factors. approaches to generating target reliability levels are used in Moses and Verma (1987) suggested target reliability levels general: (1) calibrated reliability levels that are implied in in calibrating bridge codes (i.e., AASHTO Specifications). currently used codes, and (2) cost-benefit analysis. Assuming that bridge spans of less than 100 ft are most com- The first approach is commonly used to develop reliability- mon, a T of 2.5 to 2.7 is suggested for redundant bridges, based codified design, such as LRFD. The target reliability and a T of 3.5 for nonredundant bridges. levels developed according to this approach are based on cali- Wirsching (1984) estimated the safety index, or values, brated values of implied levels of uncertainty in a currently implied by the API specifications (American Petroleum Insti- used design practice. The argument for using this approach tute, 1989) for fixed offshore structures in fatigue of tubular is that a code documents an accepted practice, and, as such, welded joints to be 2.5. He reported that this value is on the can be used as a launching point for code revision and cali- low end, because of the reference wave values. bration. Any adjustments in the implied levels should be for Madsen et al. (1986) discuss target reliability levels that the purpose of creating consistency in reliability among the were used by the National Building Code of Canada (National resulting designs when using the reliability-based code. Research Council of Canada, 1977) for hot-rolled steel struc- Using the same argument, it can be concluded that target reli- tures. The values selected were T = 4.00 for yielding in ten- ability levels used in one industry might not be fully appli- sion and flexure, T = 4.75 for compression and buckling cable to another industry. Cost-benefit analysis, the second approach to generating TABLE 13 Target reliability levels by structural type target reliability levels, is used effectively in dealing with [based on Ellingwood et al. (1980)] designs for which failures result only in economic losses and Target Reliability Structural Type consequences. Since structural failures might result in human Level (t) injury or loss of life, the use of this method might be very dif- Metal structures for buildings 3 ficult because of its need for assigning a monetary value to (dead, live, and snow loads) human life. One way to avoid the need to measure the mon- Metal structures for buildings 2.5 (dead, live, and wind loads) etary value of human life is to assign probabilities of failure Metal structures for buildings as a function of both, monitoring cost and loss of lives (see, 1.75 (dead, live, snow, and earthquake loads) e.g., Zhang et al., 2002). Metal connections for buildings 4 to 4.5 (dead, live, and snow loads) Reinforced concrete for buildings 18.104.22.168 Calibration (dead, live, and snow loads) - ductile failure 3 - brittle failure 3.5 A number of efforts for the purpose of calibrating a new Note: The t values are for structural members designed for 50 years generation of structural design codes have resulted in the of service. development of target reliability levels (i.e., safety indices, or values). The general methodology for code calibration TABLE 14 Target reliability, levels for members, based on specific reliability theories, using second-moment used by Ellingwood and Galambos (1982) reliability concepts, is outlined by Melchers (1987) and oth- Target Reliability ers. Melchers notes that frequently the information is insuf- Member, Limit State Level (t) ficient for this determination and one must make a "semi- Structural Steel intuitive" judgment in selecting target reliability, t , values. Tension member, yield 3.0 While the specific reliabilities will be a function of the Beams in flexure 2.5 Beams in shear 3.0 strength criteria needed for specific materials and load com- Column, intermediate slenderness 3.5 binations within designated structures, it is useful to have an Reinforced Concrete indication of the range of possible target reliability levels. Beam in flexure 3.0 Beam in shear 3.0 Tied column, compressive failure 3.5 Masonry, unreinforced 2.7.3 Target Reliability for Structures Wall in compression, inspected 5.0 Wall in compression, uninspected 7.5 Ellingwood et al. (1980) present ranges for reliability lev- Note: The t values are for structural members designed for 50 els for metal structures, reinforced and prestressed concrete, years of service.