Performance-based assessment of slender reinforced concrete columns typical of precast industrial buildings

Abstract

Typically, the columns of the prefabricated reinforced concrete (RC) industrial buildings and warehouses with large clear storey heights are very slender, with aspect ratios (shear span to width) larger than 10. In addition to supporting the gravity loads, the columns also provide strength, stiffness, dissipation and displacement capacity of the primary lateral-load resisting system. However, current empirical relationships that predict the non-linear response and failure mechanisms of RC columns have been developed mainly for lower aspect ratios (< 7), typical of ordinary multi-storey buildings or short-to-medium bridge piers. What makes slender columns different is their predominant flexural response, larger drifts at nominal strength and corresponding lower ductility demands, smaller ratios between the strain penetration and plastic hinge lengths to the element shear span and risk of P-Delta instability. Therefore, the direct application of analytical models available in the literature to slender columns poses a risk of overestimation of their deformation and dissipative capacity. In turns this could lead to the underestimation of their displacement demand, overall damage and collapse probability of the primary seismic-resisting and load-bearing system. In the present research extensive analysis on the non-linear response of slender columns was performed based on observed post-earthquake damage to buildings in Italy and Turkey, experimental data and numerical predictions of the failure patterns through non-linear fiber element models. The influence of the foundation flexibility and the presence of industrial floor was also investigated. The outcome is a simplified analytical methodology for the prediction of the non-linear force–deformation response and possible failure mechanisms of slender precast columns due to rebar buckling or P-Delta effects, as a fundamental step towards the seismic assessment of the global structural performance and cost-efficient retrofit solutions for precast concrete industrial buildings.

Appendices

Appendix A

1. Comparison between experimental, numerical and analytical non-linear static response of the specimens listed in Table 1. M–K and B-E indicates displacement at buckling according to Moyer and Kowalsky (2003) and Berry and Eberhard (2005), respectively.

Comparison between the experimental cyclic response and the proposed trilinear approximation

Appendix B

Analytical moment–curvature and force–displacement curves of the columns from the case-study building

See Figs. 19 and 20.

Fig. 19

Columns P14, P20 and P23, considered as cantilevers with Lc = 10.25 m

Fig. 20

Column P13, considered as cantilever with Lc = 10.25 m

The columns were considered as cantilevers. The ultimate displacement was calculated based on plastic hinge length (Equation (B.1)) according to Priestley et al. (2007). The section analysis was performed with a set of Matlab codes “Cumbia” for monotonic moment–curvature analysis of RC sections by Montejo and Kowalsky (2007) and in a simplified way using the dimensionless nominal moment M_n and yield curvature φ_y from Priestley et al. (2007). In both cases M_n and φ_y are calculated according to Equation (B.2).

$$ L_{p} = \, kL_{c} + \, L_{sp}\; {\text{where}}: k \, = \, 0.2f_{u} /f_{y} {-} \, 1 \, \le \, 0.08 L_{sp} = \, 0.0022f_{y} d_{l} $$

(B1)

$$ \varphi_{y} = \varphi_{y}^{\prime } M_{n} /M_{y}^{\prime } {\text{where}}: \varphi_{y}^{\prime } \;{\text{and}}\;M^{\prime } y\;{\text{are the first yield}}\;M_{n} \to \varepsilon_{c} \ge \, - 0.004{\text{ or }}\varepsilon_{y} \ge 0.015 $$

(B2)

The analytical force–displacement response of the cantilever columns was compared with numerical push-over curves from nonlinear FE analysis with SeismoStruct v6.5 (Seismosoft, 2013), in which the columns were modelled with inelastic force-based fiber frame elements.