Why a Curb Shouldn’t Be Kicked to the Curb: The Importance of Non-Structural Elements in Dynamic Modelling

Abstract

Occupant footfalls are often the most critical sources of floor vibration on the elevated floors of buildings, especially if rhythmic activity is expected (e.g. dancing or aerobics). Achieving reasonable vibration levels on these floors requires sufficiently stiff and massive floor structures to effectively resist the forces exerted by larger groups of people. In many cases, further vibration control is provided by Tuned Mass Dampers (TMDs). A difficulty for engineers in modelling buildings for these scenarios can be exacerbated due to the uncertainty provided by non-structural elements (e.g. non-load bearing partitions, floor toppings, curbs or railings).

In this paper, three case studies are presented of modelling structures in order to predict vibrations due to rhythmic activity. The first structure is a sports arena which features a large cantilevered balcony upon which dancing by 600 people was expected to occur. The structural design included TMDs to control these expected vibrations. Validation testing conducted once construction was complete indicated that the balcony was significantly stiffer than expected, and a complete redesign of the TMDs was required. The second structure is a long-span floor office tower that was designed with light steel trusses. Modelling predicted excessive vibration from aerobic activity on the amenity floor, which was proposed to be mitigated with TMDs. Validation testing indicated that the measured frequencies were almost 250% higher than those in the model, completely removing the need to implement TMDs. The third case study is a project consisting of two hospital towers that were nearing completion. Peer-review modelling indicated expected marginal exceedance of the required criteria, so the decision was made to measure the as-built floors. Measurements showed that frequencies were considerably higher than predicted, and extraordinarily high damping.

In all three case studies, it was concluded that non-structural elements were the cause of the large discrepancies between modelled and measured dynamic properties.