Acceleration response spectrum for predicting floor vibration due to occupant walking

Abstract

Based on 2 204 experimentally measured footfall traces from 61 test subjects, the design-oriented acceleration response spectrum was studied for calculating a floor's response due to occupant walking. Taking a rectangular floor with 42 m span as the object, the corresponding 10-second peak root-mean-square acceleration response spectrums were generated, on which a piece-wise mathematical model was proposed. The proposed response spectrum model consisted mainly three parts: the first resonant plateau ranging from 1.5 Hz to 2.5 Hz, the second resonant plateau ranging from 3.0 Hz to 5.0 Hz and the descending part ranging from 5.0 Hz to 10.0 Hz. The representing value of each plateau and the descending curve were statistically determined for different confidence levels. Furthermore, the effects of factors such as floor span, occupant's stride length, higher modes of vibration, boundary conditions and different assessment quantities on the proposed spectrum were investigated. Accordingly, the modification measure for each factor was suggested. Detailed application procedure for the proposed spectrum approach was presented and it was applied to four existing floors to predict the acceleration responses. Comparison shows that the measured acceleration responses of the four floors are generally close to the predicted results with 75% confidence level and are all lower than the predicted results with 95% confidence level. The suggested spectrum-based approach can be used for predicting a floor's response under single person's walking.