Dynamic performance of high frequency floors

Abstract

Hospitals and micro-electronics fabrication facilities require ultra-low vibration environments, and to mitigate the effects of the governing vibration source, footfalls, the floors are typically designed to have high natural frequencies so that response takes the form of a series of transients that decay rapidly between successive foot impacts. For low frequency floors evaluation of performance is relatively simple, involving simulation of resonant response in modes up to no more than 10Hz with simplifications possible through dynamic amplification factors. For high frequency floors there appears to be no simple procedure for assessment due to the complex nature of the loading time history and various difficulties with the classical normal mode analysis approach. Two approaches are in use to assess the critical peak velocity, one simple empirical formula depending on floor frequency and stiffness, and a second more recently derived empirical formula for assessing modal contributions using an effective impulse dependent on floor and walker frequency. This paper studies the reliability of the two approaches and presents some of the difficulties in attempting to obtain a realistic prediction of the performance of a high frequency floor with pedestrian loads. Because of the highly localized effects of a sharp transient load and the energy dispersion that is generally underestimated, predictions are likely to be conservative. Generally this is good, but as vibration control is a major expense, more accurate predictions could result in significant economies.