Floor vibration : dynamic properties and subjective perception

Sammanfattning: This thesis covers research in the floor vibration field. Although lightweight steel framed constructions have been in focus, many results are applicable to floors in common and to some extent also to vibrations in general. Whether a floor is treated as "good" or "bad" by the occupants is an interaction of a) the dynamic properties of the floor and of b) human vibration perception. Both these factors are covered. Two steel framed floors were tested in laboratory condition with respect to its dynamic. Modal testing was used to obtain the modal parameters, i.e. natural frequencies, damping and mode shapes. Both floors were tested for numerous set-ups where for example the effect of different supports, top layers, ceiling joists and simulated partitions were investigated. For one of the floors, subjective opinions were collected for correlation with the measured data. The research of floor dynamic has lead to a creative new method that utilizes strategically positioned small pieces of visco-elastic material to increase the damping of the floor. It was found out that this economic and proportionally easy method significantly increases the damping for some specific type of modes. The human vibration sensitivity was tested in a new developed motion simulator. Absolute threshold values for single sinusoidals were tested in the frequency range of 5-31.5 Hz. For validation of the test setup, the results were compared with data from other similar studies. The obtained thresholds agreed satisfactorily with previously reported results. More innovated tests were performed concerning multiple frequency signals of two to five discrete frequency components. The results show that a signal's composition in terms of number of frequency components and their mutual amplitude relation significantly affect the subjective perception. Therefore, it can be argued that multiple natural frequencies of a floor should be considered when it comes to design rules. This is in opposition to many floor design criteria of today since these often rely upon the fundamental frequency only. A prediction model, where the acceptance is calculated as a function of the frequency weighted total acceleration amplitude in combination with the fundamental frequency is suggested. The presented model does take into account the presence of possible multiple response frequencies.