Numerical methods to predict human induced vibrations on low frequency stairs. Part 2: Evaluation by comparing with experimental data. (December 2021)

• Record Type: Journal Article
• Authors: Andrade, Pedro; Santos, José; Escórcio, Patrícia
• Is Part Of: Structures. Volume 34(2021)
• Page Start: 1008
• Page End: 1020
• Publication Date: 2021-12
• Subjects: Ground reaction forces -- Fourier series -- Steady-state analysis -- Simplified vibration evaluation -- Vibration acceptable limits -- Experimental program; Structural engineering -- Periodicals; 624.1
• DOI: 10.1016/j.istruc.2021.08.024
• Languages: English
• ISSNs: 2352-0124

Abstract

Highlights: Measured vibrations are higher than the acceptable limits. The results of the four main numerical methods are compared with experimental data. Footfall force time histories accurately predicted vibrations. Fourier series walking models and Steady state analysis overestimated the vibrations. Adjustment factors to reduce overestimated vibrations are suggested for two methods. Abstract: Nowadays, vibration serviceability criteria are becoming the governing factor in the design of most modern staircases, because their increasing susceptibility to human induced vibrations. Although more awareness have been raised to the dynamic design of new staircases, there are still few studies found in the literature that compare the different results of numerical methods for predicting vibrations with experimental data, in order to validate the same. Hence, this paper employs the main existing numerical methods to an actual staircase with known liveness, by comparing the predicted results with the experimental data, to evaluate their accuracy when designing flexible staircases. This paper is Part 2 of a set of two papers. In Part 1, the different numerical methods are presented and details are given of how to apply them. To accomplish this, an in-situ staircase dynamic characterisation and several walking tests are performed. The measured vibrations are initially compared with different proposed acceptable limits to confirm that the vibrations exceed the limits. The different numerical methods are then employed and the predicted results are compared with the experimental results. Lastly, the main findings of this work are discussed together with those of diverse researchers who also applied one of these procedures to estimate vibrations. The results obtained showed that, with two of the numerical procedures applied (footfall force time histories and simplified vibration evaluation), it was possible to effectively predict the vibrations, while with the remaining two (Fourier series walking models and steady-state analysis), in general, overestimated responses were predicted.