Laboratory tests of steel simple torsionally unbalanced models

The objective of this work is to obtain estimations of the amplification factors α and δ used for torsion design of buildings, from experiments. For this study, simple one‐storey torsionally unbalanced (TU) steel models were considered. Models consisted of a deck supported on four columns with a selected arrangement of hinges at column ends. Two theoretical structural eccentricities (e = 0.05 and 0.15) were considered. Models were excited with a simple long‐period pendulum consisting of a hanging platform with a forced‐vibration generator on it. Eight models were tested at several excitation levels (frequencies and force magnitudes) in both ranges of behaviour: elastic and inelastic. Experiments were conducted at three frequency ratios of excitation. Registered accelerations of the pendulum platform indicate that the experimental set‐up leads to excitations that resemble narrow‐band seismic ground motions. Frame shear force estimations, based on accelerations recorded at both deck sides, indicate that torsion design factors (α and δ) depend on eccentricity. Estimations of frame shears based on measurements indicate that for normalized eccentricities e ? 0.025, the amplification α can be between 2 and 3; while δ factor resulted between 0.0 and 1.6. Copyright © 2006 John Wiley & Sons, Ltd.