Accidental Torsion in Nonlinear Response History Analysis
Licensed according to this deed.
Category
Published on
Abstract
Accidental Torsion in Nonlinear Response History Analysis
Jordan Jarrett, Reid Zimmerman, Finley Charney
Accidental torsion is ignored in many provisions for nonlinear response history analyses (NRHA), including those of FEMA P-695, FEMA P-58 and the Tall Building Initiative. Linear procedures for assessing accidental torsion exist in ASCE 7 Chapter 12 and ASCE 41 Chapter 3 where the center of mass is shifted 5% of the diaphragm dimension in each direction independently. However, studies have shown that accidental torsion can have a significant effect on the inelastic behavior of structures, and the traditional elastic methods of assessing accidental torsion may not sufficiently capture the potential inelastic torsion behavior. These effects can be especially important for structures with minimal inherent torsion. This research investigates the effects of accidental torsion on nonlinear response history analyses (NRHA), including the effects of shifts in mass and the effects of shifts in strength or stiffness.
The structure used in these analyses is a five story steel structure representative of an actual building in Berkeley, California. Its lateral force-resisting system consists of two lines of steel special moment resisting frame (SMRF) in the longitudinal direction and four lines of buckling-restrained brace frame (BRBF) in the transverse direction. Extensive nonlinear modeling is employed in Perform 3D, including localized axial-moment-moment interacting hinges for SMRF and BRBF columns, moment-rotation hinges for SMRF beam-column connections and BRBF beams outside of gusset plates, buckling-restrained brace elements and shear stress-shear strain material models for concrete walls. A suite of eleven ground motions are selected to match the characteristics of the site and scaled to a site-specific MCER target spectrum. The 3D Model was determined to meet the elastic requirements of ASCE 7 Chapter 12 as a torsionally regular structure, and minimal inherent torsion is seen during nonlinear response history analyses when accidental torsion is not included.
Multiple methods to assess accidental torsion are investigated in the research, the simplest of which is to extend the linear procedures for assessing accidental torsion that shift the center of mass by 5% of the diaphragm dimension in each direction independently. This research also uses the location of the center of rigidity to determine a method for reducing the number of necessary mass eccentricity cases from four to two. Accidental torsion from random shifts in strength and stiffness are also investigated. These analyses show that accidental torsion can have a significant effect on the inelastic behavior of this structure, especially for shifts in mass. Therefore, a check for accidental torsion within provisions for nonlinear response history analyses is warranted. This paper provides suggested language for such a provision, including exemptions that check the potential effects from accidental torsion using either nonlinear pushover analysis or individual response history analyses.