Influence of perimeter supports on the seismic response of plasterboard suspended ceiling systems

Many modern buildings across the world are equipped with suspended ceiling systems for functional and/or aesthetic requirements. Suspended ceiling are broadly classified as 1) Lay-in-tile suspended ceiling systems (LSCS); 2) Plasterboard suspended ceiling systems (PSCS). These systems are generally supported from the floor through suspenders. It may also be supported at the perimeter by the surrounding wall(s). This paper presents a study on a test set-up comprising a test frame representing a single-story building and a PSCS specimen subjected to a series of ground motions in its principal horizontal directions. The PSCS specimen was suspended through ceiling angles (suspenders) from the roof of the test frame. Properties of ceiling angles were different in the two principal horizontal directions. Ambient damping of the test frame and PSCS system were determined through hammer tests. Damping ratio was 1.5% and 1.3% for the test frame, and 6% and 8% for the PSCS specimen in two orthogonal horizontal directions. Natural period of the PSCS specimen increased with increase in shaking level: from 0.3 s to 1.0 s in one direction, and from 0.9 s to 1.0 s in other. The two directions are referred to as “long” and “short”, respectively. The change in natural period was associated with the nonlinear response of ceiling angles at their respective lower ends. Legs of the ceiling angle deformed in and out of their respective planes during the experiments. Moment-rotation behavior of the connection between ceiling angle and roof of the test frame was determined through static tests. An analytical model was developed for the analysis of the PSCS systems. The model included the behavior of connection between the ceiling angle and roof. It was assumed that the legs of the ceiling angle deform only in their plane. Natural period of the PSCS specimen obtained through the analytical model compared well with the experimental observations before first level of shaking in long direction. The natural period obtained from the analytical model in the other horizontal direction was very different from experimental values, which could be attributed to difference in the fabrication of connection between the roof of the test frame and ceiling angle during dynamic and static tests. The model was able to capture peak accelerations in the plasterboard reasonably well in the long direction. The model, however, was not able to capture the change in natural period with increase in shaking intensity, which could be due to the assumption on in-plane deformation of the legs of the ceiling angles. The model was extended to incorporate the effect of different supports at the perimeter. Support conditions included contact between the horizontal metal frame and the test frame with coefficients of friction set equal to 0.1 and 0.7, and links between test frame and horizontal metal frame. Relative displacement between the test frame and PSCS specimen decreased substantially with the introduction of a perimeter support. Peak acceleration and response spectra corresponding to the plasterboard also were considerably altered by the choice of a perimeter support in the model.