Wind Effects on Low-Rise Buildings : Database-Assisted Design vs . ASCE 705 Standard Estimates

Peak bending moments are compared for a set of steel portal frames of industrial buildings in open terrain, calculated using database-assisted design (DAD) techniques and ASCE 7-05 Standard plots. The comparisons indicate that, depending on building dimensions, the peak bending moments at the knee based on DAD techniques are generally larger by 10 % to 40 % than their counterparts based on the ASCE 7-05 plots. (In one case with a relatively steep roof slope of 26.6 ° the discrepancies exceed 90 %.) Discrepancies increase with increasing roof slope and with increasing eave height. CE Database subject headings: Aerodynamics; Buildings, low-rise; Databases; Structural design; Wind forces; Wind tunnel tests. Introduction This note compares peak bending moments in steel portal frames of industrial buildings in open terrain calculated using database-assisted design (DAD) techniques on the one hand and by using the Analytical Procedure from American Society of Civil Engineers (ASCE) Standard 7-05 (ASCE 2006, §6.5) on the other. DAD is a methodology for 1 Structural Engineer, Structural Consultants and Associates, Sugar Land, TX 77478. E-mail: coffmanb@scahouston.com 2 Research Structural Engineer, National Institute of Standards and Technology, Gaithersburg, MD 208998611 (corresponding author). E-mail: joseph.main@nist.gov 3 Research Structural Engineer, National Institute of Standards and Technology, Gaithersburg, MD 208998611. E-mail: dat.duthinh@nist.gov 4 NIST Fellow, National Institute of Standards and Technology, Gaithersburg, MD 20899-8611. E-mail: emil.simiu@nist.gov analysis and design of structures that makes direct use of pressure time histories measured in the wind tunnel (e.g., Whalen et al. 1998, Rigato et al. 2001, Simiu et al. 2003, Main and Fritz 2006). The aerodynamic database used in this study was developed by the University of Western Ontario (UWO, see Ho et al. 2005). The ASCE 7-05 Analytical Procedure entails the use of simplified coefficients, referred to in the Commentary of the Standard (§ C6.5.11 and Fig. C6-6) as “pseudo-pressure” coefficients, and based on wind tunnel data measured at UWO mostly in the 1970s (Davenport et al. 1979). The “pseudo-pressure” coefficients were developed with the aim of enveloping peak values of bending moments at the knees and ridge (see Fig. 1a), resultant vertical uplift and horizontal shear for a total of about 15 distinct building geometries. St. Pierre et al. (2005) compared bending moments, vertical uplift, and horizontal shear derived from pressures measured by Ho et al. (2005) with corresponding values computed using ASCE 7-02 plots. They noted that the responses predicted by ASCE 7-02 in many cases underestimated the responses obtained using the recent pressure measurements. These discrepancies were attributed largely to the lower turbulence intensities in the earlier experiments. Additional sources of discrepancy are that the 1970s UWO tests were conducted predominantly for wind directions in increments of 45°, as opposed to 5° in the later tests, with the number of pressure taps in the earlier tests lower by almost one order of magnitude. Also, in developing “pseudo-pressure” coefficients, the distance between frames and the structural properties of the frames had specified values, whereas in the Standard the coefficients are assumed to be valid regardless of those values. Most importantly, the values of the “pseudo-pressure” coefficients were obtained by eye, rather than by systematic and rigorous calculation. Even where the coefficients result in reasonably correct values of the bending moments at the knees and ridge, their suitability for calculating bending moments at other locations is not guaranteed. In fact comparisons by Main (2006a) of bending moments resulting from the pressure measurements of Ho et al. (2005) with those predicted by ASCE 7-05 showed that the ASCE 7-05 loads significantly under-predicted the bending moments at the “pinch” (see Fig. 1a), even for a case in which fairly good agreement was observed for the moments at the knee and the ridge. In the present study, seven buildings were selected for analysis, with dimensions listed in Table 1. All of the buildings were of gable-roofed geometry, as illustrated in Fig. 2. Four of the selected buildings had the same roof slope (θ = 14°), with eave heights H varying from 4.9 m to 12.2 m (16 ft to 40 ft). The remaining three buildings were selected to cover a range of available roof slopes in the data set from Ho et al. (2005). This represents a wider range of roof slopes than was considered in the comparisons of St. Pierre et al. (2005), who considered two roof slopes (θ = 4.8° and θ = 14°), each with four different eave heights. All buildings are low-rise structures with H < 18.3 m (60 ft), and the structural frames were designed for ASCE 7-05 Exposure C (“open country”) at a location near Miami, Florida, with a 3 s peak gust wind speed of 10m,3s V = 62.6 m/s (140 mi/h ). In all cases structural frames spaced at 7.6 m (25 ft) were considered and frame supports were assumed to be pinned. I-shaped frame cross sections with linearly varying web height were considered, and the structural analysis was performed using a linear finite element formulation. The first interior frame (see Fig. 1b) was selected for analysis, and bending moments were evaluated at the knee, the pinch, and the ridge (see Fig. 1a). For consistency, the pinch in all cases denotes a cross section located 45 % of the distance from the knee to the ridge. Analysis The bending moments corresponding to ASCE 7-05 were obtained using the Analytical Procedure (Method 2) for low-rise buildings (ASCE 2006, §6.5) with the design wind speed of 10m,3s V = 62.6 m/s (140 mi/h ) and open country terrain. The DAD bending moments were then calculated using the windPRESSURE software (Main 2006b). This software uses pressure coefficients that are referenced using the mean wind speed at eave height, with a nominal full-scale averaging time of 1 h. Such hourly averaged wind speeds at eave height can be related to 3 s peak gust wind speeds at 10 m elevation as follows: