Tests and Simple Models of RC Frame Subassemblies for Postulated Loss of Column

Abstract

The impact of the loss of an intermediate column on the most adversely affected subassembly of a multistory, multibay reinforced concrete (RC) frame was studied experimentally and analytically. Two subassemblies were considered: a monolithic one and another consisting of dry-jointed precast beams and columns, held together with the help of concentric unbonded prestressing of the beams. A uniformly distributed load was applied along the beams to realistically simulate the loading in a building with about the same geometry and loads in all floors above the failed column. The load-deflection response can be followed with a simple hand-calculation model, which takes into account the primary geometric and material nonlinearities, including arch action in the web of the beam and catenary action in the rebars and unbonded tendons. The model quantifies the difference between single-point loading of the beam, as is normally the case in other experimental campaigns, and distributed loading, as well as the contributions of the various resistance mechanisms and sensitivity to parameters. An interesting experimental finding is that a beam—even a monolithic one—does not behave as a well-integrated system of steel and concrete when it loses a column: the longitudinal reinforcement transfers the catenary tension to the beams of the adjacent bays, pulling the columns together, whereas compression from the arch action is transferred to the columns, pushing them apart. The test results are combined with those of one-way continuous slabs for simulated loss of support to an exterior beam to quantify the margins when support to a column is lost.