Vibration Transmission Through Non-Structural Partitions Between Building Floor Levels
Vibration Transmission Through Non-Structural Partitions Between Building Floor Levels
Vibration Transmission Through Non-Structural Partitions Between Building Floor Levels
net/publication/301833334
CITATIONS READS
0 28
2 authors, including:
Austin Devin
University College Dublin
11 PUBLICATIONS 31 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Austin Devin on 11 July 2018.
ABSTRACT
An assessment of the extent of vibration transmission, attributable to non-structural partitions, was made using experimental
tests and numerical simulations, for a recently constructed multi-storey reinforced concrete building. Mechanical excitation was
provided to one floor. The extent of vibration on the floor above was found to be frequency dependent and as high as 65% of
that recorded on the excited floor. A numerical model was also used to simulate the response to a person walking along a
corridor. The extent of transmission was 9% with partitions and 1.8% without them. These findings imply that the vibration level
on any given floor level will be due to excitation on that particular floor and also the not insignificant level of vibration that may
be transmitted through structural and non-structural components due to concurrent activity from floors above and below. This
issue of vibration transmission between floor levels is not addressed in any current guidance of floor vibration serviceability
assessment and requires further detailed investigation.
1. INTRODUCTION
Vibration serviceability of floor systems can be a significant issue in buildings. Traditional reinforced concrete floor systems,
with their high stiffness and damping ratios, have had an excellent track record in this regard [1]. Recently there is a trend towards
increased floor spans and slenderness, caused by a demand for faster construction times and open-plan floor layouts, which are
more likely to exhibit vibration serviceability issues due to their slenderness and lower energy dissipation capability [2-4].
It has also been illustrated that non-structural elements can have significant effects on the dynamic response of floor systems of
buildings; evidence of such effects was observed in determining the dynamic parameters of both slender floor systems [5-6] and
traditional concrete floor systems [7]. Exclusion of non-structural elements has been shown to result in discrepancies between a
numerical model of a building and its actual structural response. Miskovic et al. [6] analysed the effect of full-height partitions
using forced vibration testing. Ventura et al. [8] used ambient vibration results to update a Finite Element (FE) model of a
building including non-structural elements. Li et al. [9] highlighted that measured natural frequencies of a 79 storey building
were higher than those calculated from the FE model, concluding that this difference was due to the presence of non-structural
elements. No significant research, to the authors’ knowledge, has been undertaken to date on the extent of vibration transmission
from one floor level to the next due to these non-structural partitions.
This paper describes a combined experimental and numerical investigation in to the extent of vibration transmission between
one floor and the next. In a series of experimental tests mechanical excitation was provided on one floor level of a recently
constructed reinforced concrete building, a traditional column and slab system, with non-structural partitions forming corridors
and various working spaces on each floor level. Responses were recorded on the excited floor level and the one above to measure
the extent of vibration transmission during these tests. It was not possible to repeat these tests without partitions in situ and hence
a validated numerical model [7] was used to compare the response on two floors, using models with and without partitions, due
to a person walking along a corridor length in the bulding. The aim of the simulation was to attempt to determine the contribution,
if any, of the partitions to vibration transmission between floors.
The Charles Institute on the University College Dublin (UCD) campus in Ireland is a four storey reinforced concrete frame office
building. Structurally, the frame consists of two-way spanning flat slabs, 0.3m thick, supported by 0.4m square columns with a
maximum bay size of 7.5m × 6.6m. The lateral load resisting system is made up of a number of reinforced concrete stairwells,
lift cores and service ducts with wall thicknesses of 0.2m. Each floor level is divided into state-of-the-art laboratories and office
accommodation using lightweight non-structural partitions. These partitions consist of two layers of gypsum boards, 25mm total
thickness, attached to both sides of lightweight metal studs at 400mm centres which house an insulation material. Exterior wall
cladding consists of large polished Chinese black basalt panels 40mm thick supported on galvanised steel rectangular sections
fixed to the concrete slab above and below at 400mm centres. The structure of each floor is identical with only the layout of
internal partitions varying from floor to floor. The layouts of partitions, for the two floor levels considered, are shown in Fig. 1a
and 1b. In both cases the solid black lines indicate the partition layout on the floors below.
Walking trajectory
Shaker at base of
partition
Fig. 1a First floor internal partition layout. Fig. 1b Second floor internal partition layout
The objective of the mechanical excitation tests was to determine the vibration response on floor level 2 due to excitation on
floor level one. An APS Dynamics electrodynamic shaker, with a capacity of 500 N, was used to provide vertical excitation on
floor level 1, while Endevco 7754-1000 piezoelectric accelerometers (at quarter points and mid-way along the partition at its
base of partition and immediately above it on floor level above) were used for response measurements. A sweeping 'chirp' force
was applied to Floor Level 1 at the base of, and mid-way along, an internal partition (‘A’ in Fig. 1a). The frequency range for
the force was 5 Hz to 40 Hz, sweeping a rate of 0.63 Hz per second; the maximum applied force was 480 N, reducing to 100 N
over the length of the ‘chirp’. The resulting vibration response of Floor Levels 1 and 2 mid-way along the partition are shown in
Fig. 2. As expected the highest response is recorded on the excited floor level but the level of transmission to floor level 2 is not
insignificant. The acceleration time histories were used to calculate a percentage of vibration transmission; the percentage of
vibration transmission being defined as the ratio of the instantaneous vibration level at Floor Level 2 to that at Floor Level 1.
This percentage of vibration transmission is plotted against time in Figure 3 for each of the three sensor pairs along the length of
the partition. The extent of vibration transmission varied between approximately 10% and 65% over the variation of the chirp
test – the average level of transmission was of the order of 32%.
Fig. 2 Acceleration time histories mid-way along partition on floor levels 1 & 2
Ideally the mechanical excitation tests would have been repeated on the bare frame system in the absence of partitions but it
was not feasible to remove all partitions. Thus while it is obvious from the test data that there is transmission of vibrations from
one floor level to the next it cannot be confirmed categorically whether, and to what extent, the partitions contribute to this
transmission. Thus, the response of validated numerical models, discussed subsequently, in which partitions are included and
also removed, were also studied.
4. VIBRATION TRANSMISSION DUE TO SIMULATED WALKING
During the process of quantifying the effect of non-structural partitions on the modal characteristics of the Charles Institute
building [7,10-11] validated finite element models of both the bare frame (columns, slabs and shear walls only) and completed
structure (additional inclusion of including non-structural partitions and external façade cladding units, both modelled using
equivalent vertical springs) were developed. The representation of the bare frame in each of these models is identical and thus
provides the opportunity to evaluate vibration transmission from one floor level to the next numerically. The finite element
meshes, where the nominal element size on each floor level is of the order of 0.4m x 0.4m, are shown in Figs. 4a and 4b.
For the purposes of this numerical evaluation the responses due to a person walking along a section of floor (indicated as
‘walking trajectory’ with an arrow in Fig.1a) on floor level one are evaluated. The pedestrian load applied, as a function of time
in Newtons, after [12] is given by Eq.1. It describes the temporal variation in load due to a pedestrian of mass 80kg walking at
a pacing rate of 1.8Hz with a first harmonic factor of 0.4.
For the purposes of considering vibration transmission the responses at a point mid-way along the walking trajectory and at a
point located immediately above this, on floor level two, were considered. A transient dynamic analysis was performed using
a time step of approximately 0.015 seconds so as to adequately capture the load dependency on the pacing frequency of 1.8Hz.
The analysis duration was 7 seconds. For each time step the force F(t) and position along the walking path was determined.
The load was then shared pro-rata to the two closest node points along the path and ramped from prior values. Because of the
un-even nature of the mesh (node to node distances were nominally 0.4m but not exactly 0.4m) this led to a unsteady high
frequency content in the acceleration response which made direct comparison of effects difficult using acceleration responses
– displacement responses, at the two reference points, were thus employed for evaluating comparative accelerations. These
displacement (in the vertical direction) responses are plotted in Figure 5.
The response at each reference point is an oscillatory response at 1.8Hz due to the walking load model pacing frequency applied.
The peak acceleration at the reference point occurs when the load is at the reference point and is determined from the maximum
peak to peak deflection divided by two and multiplied by the response frequency squared in radians per second. The peak
accelerations, so calculated, are listed in Table 1.
The model without partitions is characterised by higher accelerations on the load floor level but relatively little transmission of
vibration (ratio of peak floor level 2 acceleration to floor level 1 acceleration) to the floor above. In the absence of partitions the
level of transmission is 1.8% compared to 9% when partitions are included.
With Partitions Without Partitions
5. DISCUSSION OF RESULTS
The experimental and numerical investigations into vibration transmission between floor levels discussed above seek to explore
the contribution of non-structural partitions to such transmission. Both sets of investigations show that excitation on a given
floor level can be expected to produce a vibration response on the floor above.
In the case of the numerical simulations the presence of the partitions results in a stiffer system which is consistent with prior
experimental modal studies [7] which showed that nominally identical floors in the absence of partitions had increased
frequencies and altered modes of vibration once the partitions were added. Hence with partitions included although the vibration
levels (Fig.5 and Table 1) identified on the floor along which walking was simulated are lower the extent of transmission of
vibration levels to the floor above is higher – the upper floor vibration level is predicted to be of the order of 9% of those on
the excited floor compared to only 1.8% in the absence of the partitions. This constitutes a fivefold increase in transmission
from one floor to the next due to the presence of partitions.
While it was not possible to undertake the experimental test programme with and without partitions the extent of transmission
was at least as pronounced, being in the range of 10% - 65% over the duration of the test. Vibration transmission of the order
of 10% is considered significant and implies that the actual vibration response on a given floor level is likely to be due to the
activity on that floor level itself and also to some not insignificant level of transmission from the floors below. This is currently
not acknowledged in any code of practice for floor vibration serviceability assessment and is in the authors’ opinions an area
of research that requires further investigation.
6. CONCLUSIONS
Experimental and numerical studies have shown that there is transmission of vibration from one floor to the next. The extent
of transmission along a partition line was measured to be in the range of 10% - 65% for a recently constructed reinforced
concrete building with different partition layouts on each floor level.
Numerical models of the test structure indicate that the extent of transmission is increased, a fivefold increase in this case, due
to the presence of partitions. It follows that building floor vibration serviceability assessment should account for this
transmission. Currently this in not the case and further research work, preferably experimentally based, on full scale structures
with and without internal partitions, is required.
ACKNOWLEDGEMENTS
The authors wish to express their gratitude to i) the Irish Research Council for Science, Engineering & Technology for their
financial support, ii) the Vibration Engineering Research Section of Sheffield University for use of their equipment at the time
the testing was undertaken, and iii) the Buildings Office at University College Dublin (UCD) for access to the Charles Institute
at UCD.
REFERENCES
[1] Pavic, A. et al., 2001. Critical review of guidelines for checking vibration serviceability of post-tensioned concrete
floors. Cement and Concrete Composites, 23(1), pp.21–31.
[2] Pavic, A., Miskovic, Z. & Reynolds, P., 2007. Modal Testing and Finite Element model updating of a lively open-
plan composite building floor. Journal of Structural Engineering, 133(4), pp.550–558.
[3] Pavic, A., Widjaja, T. & Reynolds, P., 2002. The use of modal testing and FE model updating to investigate vibration
transmission between two nominally identical building floors. In Proc., Int. Conf. on Structural Dynamics Modeling-
Test, Analysis, Correlation and Validation. Instituto Superior Tecnico Lisbon, Portugal, pp. 347–355.
[4] Nyawako, D., Reynolds, P. & Hudson, M., 2013. Findings with AVC design for mitigation of human induced
vibrations in office floors. In Topics of Dynamics in Civil Structures, Volume 4. Springer, New York, pp. 37–44.
[5] Reynolds, P., 2000. The effects of raised access flooring on the vibrational performance of long-span concrete floors.
PhD Thesis: University of Sheffield.
[6] Miskovic, Z., Pavic, A. & Reynolds, P., 2009. Effects of full-height nonstructural partitions on modal properties of
two nominally identical building floors. Canadian Journal of Civil Engineering, 36(7), pp.1121–1132.
[7] Devin, A., Fanning, P.J. and Pavic, A. (2015) 'Modelling effect of non-structural partitions on floor modal properties'.
Engineering Structures, 91 :58-69.
[8] Ventura, C.E., Lord, J.-F. & Simpson, R.D., 2002. Effective use of ambient vibration measurements for modal
updating of a 48 storey building in Vancouver, Canada. In International Conference on Structural Dynamics Modeling
– Test, Analysis, Correlation and Validation. Madeira Island, Portugal.
[9] Li, Q.S. et al., 2002. Field measurements of amplitude-dependent damping in a 79-storey tall building and its effects
on the structural dynamic responses. The structural design of tall buildings, 11(2), pp.129–153.
[10] Devin, A. & Fanning, P.J., 2012. The Evolving Dynamic Response of a Four Storey Reinforced Concrete Structure
during Construction. Shock and Vibration, 19(5), pp.1051–1059.
[11] Devin, A., 2013. The Effect of Non-structural elements on the Vertical Dynamic Response of Floors. PhD Thesis.
School of Civil, Structural and Environmental Engineering, University College Dublin.
[12] Bachmann, H. & Ammann Vibrations in Structures Induced By Man and Machines, International Association for
Bridge and Structural Engineering, Structural Engineering Document, Zurich, 1987.