Abstract

US building design code,ASCE 7-16[1]defines the same return periods concept for seismic and wind loads based on the probability of collapse of buildings.The return periods of risk targeted maximum considered earthquake(MCER)in ASCE 7-16 are from 1 000 to 3 000 years(National Institute Building Sciences Building Seismic Safety Council,2015).The design basis earthquake(DBE)is 2/3 of MCER;however,considering importance factor of 1.0 to 1.5,converted return period of DBE is up to 3 000 years.Wind loads in ASCE 7-16 use return periods from 300 to 3 000 years.Although the return periods of seismic and wind loads are similar,there is a large difference in target performance of structure.An elastic design is used for wind loads,while inelastic behavior is permitted under seismic loads.An elastic design is carried out using seismic loads reduced by response modification coefficient(R factor)considering ductility of system.The difference of target performance of two loads causes problems in the design of high-rise buildings.

The design seismic load can be reduced by R factor,because buildings are assumed to have sufficient ductility with a desirable yield mechanism.The design wind load can be larger than reduced seismic load in the high-rise buildings with aspect ratio larger than 3[2],and the elastic design is required.To satisfy elastic design for wind load,the size of horizontal members is increased.It results in yielding of vertical members and joints rather than horizontal members like beams and braces under strong earthquakes.The yield of vertical members and joints prior to horizontal members leads to a brittle system as shown in Figure 1.The initial stiffness and strength of system increase,but the ductility decreases.Thus,the R factor with assumed sufficient ductility in the seismic design becomes invalid.Because the real seismic load without R factor is still larger than wind load,the safety of structure cannot be guaranteed.Due to this issue,the necessity of inelastic wind design,so-called performance-based wind design(PBWD)is recognized.

Figure 1　Simplified yield mechanism issue of a frame in high-rise buildings caused by elastic wind design

Although no detailed implementation method has been proposed so far,ASCE recently published Prestandard for Performance-Based Wind Design[3].It proposed limited inelastic performance objectives for wind load with return periods from 700 to 3 000 years.

To carry out PBWD,verification of performance by nonlinear analysis is necessary,and before that an initial design is required.The R factor is used in initial seismic design(prior to performance-based seismic design),but the whole wind load cannot be reduced like seismic design.Wind load is composed of mean,background,and resonant components.The mean and background components are essentially quasi-static loads,so it makes sense that a structure cannot be designed to be inelastic under these components.It appears more reasonable that the response modification factor for wind load(RW)would be applied to the resonant component only[4].Because the duration of wind load is much longer than that of seismic load,the RW factor should be less than R factor considering fatigue(i.e.1.5 to 2.0).Because the portion of resonant component in the along-wind load increases for increased building height and the across and torsional wind loads are mostly composed of resonant component[5],the design wind load can be ef fectively reduced by an RW factor of 1.5 or 2.0.By reducing initial design wind load,overdesign of horizontal members can be precluded.After the initial design,exact dynamic performance under original wind loads should be verified by nonlinear time history analysis.Time histories of wind load are obtained from wind tunnel tests or generated from power spectral density functions.

As buildings become higher,PBWD is expected to be necessary to secure the structural performance and efficiency.