Engineering Structures 33 (2011) 24852494

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Control of vibrations induced by people walking on large span composite floor decks
Wendell D. Varela a, , Ronaldo C. Battista b
a

Faculdade de Arquitetura e Urbanismo, Universidade Federal do Rio de Janeiro, Departamento de Estruturas, Av Pedro Calmon, 550, Prdio da Reitoria/FAU, Cidade Universitria, Rio de Janeiro, RJ, CEP 21941-901, Brazil
b

Instituto COPPE, Universidade Federal do Rio de Janeiro, Engenharia Civil, C. Postal 68506, CEP: 21945-970, Rio de Janeiro, RJ, Brazil

article

info

abstract
The low damping properties of lightweight large span floor decks composed of a reinforced concrete slab on top of a steel space frame structure may lead to undesirable dynamic responses, even to ordinary human actions such as walking. This problem was investigated through laboratory tests performed on a 1:1 scale prototype of a composite floor deck structure. Experimental measurements were taken for the structure subjected to several dynamic human loads, especially those produced by the random walking of people. To compensate for the lack of damping, a passive control system was designed and installed in the composite structure prototype. The performance of the mechanical control devices was evaluated by means of straight comparisons between the experimental acceleration amplitudes obtained for the controlled and uncontrolled structure subjected to similar dynamic forces produced by one or more persons walking. The most relevant results for both time and frequency responses are presented and used to argue that small and low cost passive control devices can already be included in the design stage of a smart structure as effective accessories to substantially reduce vibrations induced by people in low damped large span composite floor decks. 2011 Elsevier Ltd. All rights reserved.

Article history: Received 17 March 2010 Received in revised form 13 April 2011 Accepted 15 April 2011 Available online 1 June 2011 Keywords: Dynamic control Composite structures Floor decks Human loads Vibration

1. Introduction Many excitation sources may generate excessive vibrations in low damped lightweight large span floor decks, which may cause discomfort to the users of residential, commercial, or public buildings. Typical examples of these vibration sources are traffic of heavy vehicles on neighboring roads, large machinery in neighboring constructions sites, equipment installed in buildings, and human activities on floor decks, such as walking, running, and jumping. Among these vibration sources, those produced by human activities are the most common, and walking is a daily activity everywhere. The number of cases of floor deck structures that present human-induced vibration problems has grown with the increasing number of building constructions with composite floor decks spans that are larger and more slender. Vibration problems induced by human walking have been observed for almost two centuries; in

Corresponding author. Tel.: +55 21 25981682; fax: +55 21 2598 1890. E-mail addresses: wendelldv@gmail.com (W.D. Varela), battista@coc.ufrj.br (R.C. Battista). 0141-0296/$ see front matter 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.engstruct.2011.04.021

1828, Tredgold [1] apud [2] suggested that longer girders should be built with a greater cross-section height to avoid everything shaking in a room when someone walks on the floor. In this same direction, structural stiffening has been largely favored by engineers as a practical design solution and as a remedial measure to reduce vibrations in existing structures; however, in certain cases [3], this may lead to non-practical dimensions of structural components or to a cost-benefit trap. In the problem addressed in this study, people are simultaneously vibration sources and displeased users, and sometimes they may act as energy dissipation devices [4,5]. Although it is possible to lessen the intensity of the human vibration source by covering the floor with layers of fabric and rubber-like materials, more efficient and lasting solutions are achieved by increasing the damping properties of the structure, by installing dynamic control devices, or, if there is a cost benefit, by stiffening the structure as in common practice. Whenever allowed by the inner architecture layout, a cheap solution may be achieved by installing partition walls along with struts topped with rubber pads compressed against two floor slabs, combining the required increase in structural stiffness and damping [2,6].

2486

W.D. Varela, R.C. Battista / Engineering Structures 33 (2011) 24852494

Fig. 1. Application of multiple TMDs on prefabricated slabs [4]. (a) Full view. (b) Detailed view of a TMD.

Smart concretes with special damping components [7,8] or a sandwich reinforced concrete slab with a thin inner layer of viscoelastic material [9,10] can be possible solutions. A substantial reduction of vibrations induced by human walking on floor slabs may otherwise be achieved by properly designing and installing tuned mass dampers (TMDs), which are efficient, low cost, and low maintenance mechanical devices that can be installed without interrupting operational or human activities in the building. Another advantage that TMDs have over other alternative solutions is their versatility; they can be designed in many different shapes and sizes as needed, and as required to accommodate the architectural aspects and space limitations. A more advanced control solution can be sought by using active mass dampers (AMDs), which in general lead to greater reductions in vibration amplitudes in relation to those achieved with TMDs [11] because of the faster response of an active control system and because of the capacity of such a system to operate in a band of frequencies, which is a clear advantage considering the random characteristics of human walking. However, they present some practical drawbacks: a source of electric energy is needed for operation; special maintenance is required, and they are somewhat more expensive than their passive counterparts because they are composed of several electric and electronic components as well as smaller mechanical devices. In the present work, it is shown that the problem of excessive vibration amplitudes produced by the random walking of one or more persons on a composite floor deck structure may be properly solved by using TMD-like passive control devices. 2. Mechanical devices to attenuate vibrations induced by people on structures Subsidiary dynamic devices to reduce the amplitudes of motion in mechanical systems were first conceived by the German engineer Frahm [12] in 1911, and later applied by German naval architects to control the rolling motion of ships. Ormondroyd and Den Hartog, in 1928 [13,14], derived the mathematical equations for a two-degree-of-freedom controlled mechanical system, in which the controlling device was called a dynamic absorber. In civil engineering applications, one of the first studies was conducted by Lenzen [3], who very briefly indicated the possibility of adopting control devices to damp vibration induced by people on floor decks; these devices were then called tuned mass dampers (TMDs). Many other studies on the practical application of dynamic attenuators in large civil engineering structures have been conducted by researchers in various countries. It has been demonstrated theoretically and experimentally [1526] that the TMD efficiency depends on the ratios between its frequency and damping to the fundamental frequency and modal damping

Fig. 2. Large span composite floor deck structure used for human walking tests.

of the structure, respectively. Multiple synchronized dynamic attenuators, MSDAs, can significantly reduce floor slab vibration amplitudes in one or multiple closed spaced modes of vibration. Battista and Varela [2729] designed an MSDA system tuned to four different frequencies to reduce vibrations induced by humans walking on a continuous four-panel composite floor deck (Fig. 1(a) and (b)). The conceptual design of these attenuators followed that of a simple cantilevered beam with a lump of mass at its tip. For any of the attenuators, frequency tuning was done by sliding the lumped mass along the beams length. The mechanical damping inherent to the bolted pieces of the attenuators suffices for their good performances. The attenuators were located at points of maximum amplitudes of the dominant vibration mode shapes. Over 80% reduction of vibration amplitudes was achieved, contradicting current expectations and past concepts [15]. 3. Experimental program The vibration problem of interest was investigated [4,28] through laboratory tests performed on a 1:1 scale prototype of a 9.2 9.2 m square composite floor deck (Fig. 2). The prototype structure is composed of a reinforced concrete slab with 13 cm thickness on top of a steel space truss formed by pyramidal modules with a 1.15 1.15 m square base (Fig. 3). The floor deck is a reinforced concrete slab on a corrugated metal sheet with 0.65 mm thickness. The concrete has elastic modulus 30.5 GPa and specific weight 24.5 kN/m3 . The connections of the steel shapes are all bolted. The total mass of the structure is 35 t. This large span structure was conceived by Souza and Battista [30,31] to be employed as floor decks in shopping center buildings. The experimental program was divided into two phases. In the first phase, free vibration tests were carried out to obtain some of the dynamic characteristics of the structure, such as the first three natural frequencies as well as the corresponding mode shapes and

W.D. Varela, R.C. Battista / Engineering Structures 33 (2011) 24852494

2487

Fig. 3. Cross-section of the large span composite floor deck structure.

modal damping ratios. These tests additionally served as a basis for designing the TMD-like mechanical control devices. In the second phase, human walking tests were performed on the uncontrolled structure and on the controlled structure to evaluate the efficiency of the control devices in reducing the vibration amplitudes. 4. Free vibration tests of the structure Free vibration tests were performed by applying heel impacts on the deck structure. Accelerometers were installed at characteristic points (Fig. 4) to pick up the first three mode shapes of the structure (Fig. 5(a)(c)). Table 1 presents the experimental frequencies and damping ratios associated with the first three vibration mode shapes. The mode shapes depicted in this table and their related modal masses were obtained with an experimentally calibrated refined finite element numerical model. The finite element
Fig. 5. The first three experimental vibration mode shapes drawn with the instantaneous acceleration amplitudes measured at the points indicated in Fig. 4.

model (FEM) using the actual dimensions and mechanical properties of the materials led to a good match between the experimental and numerical results, i.e. no special adjustment of the numerical model was needed. The values of the natural vibration frequencies of the structure were obtained by applying a fast Fourier transform (FFT) on the acceleration versus time responses. The modal damping

Fig. 4. Instrumentation plan for tests indicating locations of the accelerometers and points of heel impacts.

2488

W.D. Varela, R.C. Battista / Engineering Structures 33 (2011) 24852494

Table 1 Dynamic characteristics of the tested floor deck structure prototype. Mode number Frequency (Hz) Damping ratio Modal mass 1 7.57 (0.03) 0.25% 20 t 2 12.21 (0.03) 0.64% 10 t 3 12.42 (0.03) 0.66% 10 t

Vibration mode shape

Fig. 6. TMDs used in the tests. (a) Schematic of the TMDs installed in the floor deck structure. (b) Pieces used in the fabrication of one of the TMDs.

ratios were obtained by the log decrement method applied to the frequency band filtered acceleration versus time responses. As the damping ratios were of quite low values, a different estimation method was used. The results obtained with Ibrahims method [32] confirmed the values obtained with the log decrement method. The amplitudes of the first vibration mode of the structure (7.57 Hz) can be greatly amplified by resonance with the fourth harmonics (6.49.6 Hz) of human walking excitation forces whose fundamental frequencies fall in the range 1.62.4 Hz. Furthermore, the very low damping ratio (0.25%) of the first vibration mode contributes substantially to the amplification of the response amplitudes, which may cause discomfort to the users of buildings constructed with composite floor decks. 5. Mechanical devices for dynamic control The preliminary human walking tests on the floor deck structure showed the need for attenuating the vibration amplitudes of the dynamic responses to improve human comfort. To compensate for the lack of damping of the composite floor deck, a passive control system was designed and installed under the reinforced concrete slab at the center of the prototype structure. The design of the TMD-like mechanical control devices was based on the experimental results obtained for the uncontrolled structure and parametric studies following the authors previous experience in design works [2729,33]. Fig. 6(a) shows a schematic of the TMDs installed in the structure. The mobile mass of the TMD is composed by four circular steel plates sitting on three compressed helicoidal springs. Fig. 6(b) shows the many pieces used in the fabrication of one of these control devices. Fig. 7(a) shows the two TMDs installed under the slab, and Fig. 7(b) shows a detailed view of one of the TMDs. Two TMDs were chosen, to make them lighter, and to make them easier to install. For larger mass ratios (i.e. the ratio between the total mass of the TMDs and the modal mass of the structure related to its fundamental vibration mode), a larger number of TMDs or a system of multiple synchronized dynamic attenuators [33] is

recommended. An added advantage of multiple TMDs is related to the operational efficiency: although very robust, if eventually one breaks the others will continue to perform their work with a small loss of efficiency. The two TMDs were tuned to frequencies fTMD close to the first natural frequency f1 , resulting in a frequency ratio fTMD /f1 0.97. A frequency ratio 0.950.97 allows for changes in modal mass or stiffness (due to added dead loads or structural alterations), which may occur in the service life of the structure. The total mass of each TMD (mobile mass plus the other parts) is approximately 55 kg. The effective total mass of the two TMDs is approximately 94 kg (each with a mobile mass of 47 kg), which corresponds to 0.47% of the modal mass (20 t) related to the first vibration mode or 0.27% of the total mass (35 t) of the whole composite structure. A mass ratio of 0.21% is, in general, adequate for obtaining substantial reductions of vibration amplitudes. After the fabrication, the TMDs were bench tested to check their mechanical behavior and to verify if the required dynamic characteristics were fully attained. The bench tests of the TMDs were performed by exciting the mobile mass manually; this was instrumented with a micro-accelerometer. The damping coefficient was then estimated by the logarithm decrement technique applied to the free vibration response signal. The hysteretic damping of the TMDs was inherent to their own mechanism and some friction between their moving masses and guiding axles. The values of the damping ratios obtained in the bench tests fell in the range 11.5%. Despite this low damping, the motion amplitude of the mobile mass is very small in the working frequency of the TMDs. The springs shortening equals this amplitude, which yields an even smaller variation in the pitch of the coils. This range of damping allowed for a good performance of the TMDs. It is worthwhile noting that a too large damping coefficient may lock the TMDs mobile mass in the structure. 6. Human walking tests A series of tests was performed on the controlled and uncontrolled floor slab prototype model under excitation produced by one and several persons walking (see Fig. 8(a)(c)).

W.D. Varela, R.C. Battista / Engineering Structures 33 (2011) 24852494

2489

Fig. 7. TMDs installed on the floor deck structure for controlled tests. (a) Full view. (b) Detailed view of a TMD instrumented with an accelerometer.

be observed that the frequencies and lengths of the volunteers steps were spontaneous, i.e., without the aid of any visual or audio stimulations. Tests A and C lasted 180 s and Test B lasted 60 s. The duration of each test represents a significant sample of the randomness of the human walking characteristics. The statistical parameters for walking extracted from an extensive editing work of video records and careful frame-toframe visual observation (30 frames per second resolution) for both uncontrolled and controlled tests (Table 2) were shown to be similar. Furthermore, it can be noted in Figs. 9(a)(c), and 10(a) and (b) that the walking trajectory prints and frequency histograms are also similar for related controlled and uncontrolled tests. These trajectory prints and histograms were constructed as follows: for each person, and for each test, the points where people touched the floor with the heel were marked on grids of orthogonal lines, and the corresponding elapsed time of the test were registered. Then, these marked grids were scanned and passed through image recognition software. The coordinates of the walking path were obtained by taking the known coordinates of the floor deck boundaries. The walking frequencies could also be obtained from the registered time intervals between the marked points. 7. Experimental results The performance of the control system was evaluated by means of straightforward comparisons of the measured acceleration amplitudes in both controlled and uncontrolled structures under similar dynamic forces. In other words, each test was performed twice: first for the uncontrolled structure in which the TMDs were locked, and next for the controlled structure, both subjected to dynamic forces produced by the same people walking with almost the same trajectories and step frequencies. The similarity between the tests performed sequentially on controlled and uncontrolled structures was proven by comparative analysis of the resulting human walking statistical parameters and trajectory prints. Comparisons between the controlled and uncontrolled dynamic responses are presented in both the time and frequency domains. The efficiency of the mechanical control devices in reducing the amplitudes of vibrations induced by human walking in each type of test was evaluated according to three different criteria: (i) direct comparison of the percentage of reduction in vibration amplitudes; (ii) volunteers perception to vibrations; and (iii) the human perception criterion as recommended by ISO 2631-2 [34], and the acceptance criteria for human comfort recommended by SCI [35] and AISC [2]. Fig. 11(a) and (b) show the dynamic responses in terms of the vertical acceleration at the center of the composite deck for the uncontrolled and controlled structures, respectively, in Test A. These figures also show the TMD responses in two distinct conditions: (a) locked, and turned into small added masses

Fig. 8. Walking tests performed on the prototype composite floor deck structure. (a) Test A: one person walking. (b) Test B: six persons walking side by side. (c) Test C: six persons walking randomly.

Several tests were performed, and the most relevant results of the experimental measurements are summarized herein for the three most representative tests on the structure. (i) Test A: one person (0.73 kN weight) walking in a predefined trajectory with an approximately constant step frequency in quasi-resonance with a subharmonic of the fundamental frequency of the structure (Fig. 8(a)); (ii) Test B: six persons walking (average weight 0.77 kN) side by side with closely synchronized steps (Fig. 8(b)); and (iii) Test C: six persons walking (average weight 0.77 kN) randomly on the slab and at random step frequencies (Fig. 8(c)). It should

2490

W.D. Varela, R.C. Battista / Engineering Structures 33 (2011) 24852494

Fig. 9. Trajectory prints of people walking during tests. (a) Test A: one person walking. (b) Test B: six persons walking side by side. (c) Test C: six persons walking randomly (trajectory shown for just one person). Table 2 Statistical parameters obtained from the human walking tests. Test Walking frequency (Hz) Average A B C Uncontrolled Controlled Uncontrolled Controlled Uncontrolled Controlled 1.87 1.85 1.77 1.74 1.74 1.71 Standard deviation 0.07 0.07 0.06 0.06 0.08 0.08 Step length (m) Average 0.79 0.79 0.72 0.70 Standard deviation 0.07 0.07 0.07 0.07 Velocity (m/s) Average 1.48 1.47 1.24 1.20 Standard deviation 0.14 0.13 0.14 0.13

Fig. 10. Walking frequency histograms. (a) Test A: one person walking. (b) Test C: six persons walking randomly.

clamped to the structure (the TMDs and the structure present the same acceleration amplitudes at the same location); (b) free, and operating as proper dynamic control devices (the TMDs and the structure present different acceleration amplitudes at the same location). The maximum relative displacement amplitude of TMD mass with respect to the structure (estimated by double

integration of the acceleration signals) was attained for Test A, and reached 0.73 mm. A great reduction in the vibration amplitudes of the structure was achieved with the two TMDs, as can be observed in Fig. 11(a) and (b), through direct comparison between the controlled (b) and uncontrolled (a) responses of the structure. This reduction of vibration amplitudes is explained by the TMD operation as follows: the TMD frequency, which was tuned near the first natural frequency of the structure, and their appropriate location on it (at a point of greatest amplitude in the first vibration mode shape) cause the acceleration of the steel mobile masses of the TMDs, generating inertia forces that are in opposition to the inertia forces produced by the external action on the structure. Figs. 1217 show comparisons between the measured uncontrolled and controlled dynamic responses (in both the time and frequency domains) of the structure for Test A to Test C given in terms of the vertical accelerations at the deck center. It can be noticed in the experimental frequency responses (Figs. 1517) that the acceleration amplitudes of the first vibration mode (7.57 Hz) show large reductions (4784%) in the controlled structure compared to the amplitudes of the uncontrolled structure. In Figs. 1517, the synchronized control devices (TMDs) introduced an additional degree of freedom into the structural system that resulted in two new eigenfrequencies of the controlled system around the first vibration frequency (7.57 Hz) of the uncontrolled structure: one slightly smaller (fTMD = 7.32 Hz) and one slightly higher (f1 = 7.81 Hz). The greatest reductions in vibration amplitudes were found in Test A for only one person walking, which can be explained by the quasi-resonance phenomena produced by a component of the excitation force (the fourth harmonic of the walking load; 4 1.87 Hz = 7.48 Hz) on the first vibration mode of the uncontrolled structure (7.51 Hz for the locked TMDs), as shown in Figs. 12 and 15(a). These results demonstrate that the efficiency of the passive control system is greater when one harmonic component of the walking loading has a frequency close to the frequency of one of the structural vibration modes to which the TMDs are tuned, which dominates the dynamic response of the structure.

W.D. Varela, R.C. Battista / Engineering Structures 33 (2011) 24852494

2491

Fig. 11. Structure and TMD time responses for Test A. (a) TMDs locked. (b) TMDs unlocked.

Fig. 12. Uncontrolled and controlled time responses of the structure for Test A. (a) Structure responses. (b) Enlargement of the time response: resonance of the fourth harmonic of the walking force with first vibration mode frequency.

Fig. 13. Time response for Test B.

Fig. 14. Time response for Test C.

In Fig. 15(b), the uncontrolled and the controlled responses in Test A have close peak amplitudes in the first three harmonics of the walking force in a range of frequencies far from resonance to the vibrational frequencies of the structure. This closeness demonstrates that both the uncontrolled and the controlled tests were performed for walking excitations with very similar dynamic characteristics, which validates all the correlations made between the controlled and uncontrolled measured responses. As compared to the results of Test A, those obtained in Tests B and C show, for the controlled structure, smaller but yet important reductions of the response amplitudes in the first vibration mode. This is explained as follows: the response amplitude of the structure at any instant of time is composed of the amplitudes of multiple vibration modes in varied frequencies. On the other hand, both TMDs were tuned to the first mode frequency. As a

consequence, no reduction of response amplitudes is expected for other higher frequency vibration modes. In Tests B and C, people walked without synchronism and at walking frequency ratios apart from resonance with the fundamental vibration mode as occurred in Test A. That is why the observed reductions are greater in Test A, and that is why the observed reductions of vibration amplitudes in the non-filtered acceleration response signals are very different from the root mean square (rms) filtered acceleration responses in Tests B and C. It should be emphasized that the evaluation of the efficiency of the TMDs was made with the rms values of the signal filtered to a short bandwidth around the first vibration mode frequency. An interesting fact is that the amplitudes of vibrations for one person walking in Test A reached values close to that of six persons walking side by side in Test B and greater values than for six

2492

W.D. Varela, R.C. Battista / Engineering Structures 33 (2011) 24852494

Fig. 15. Frequency response for Test A. (a) Frequency response. (b) Enlargement of the frequency response.

Fig. 16. Frequency response for Test B.

Fig. 17. Frequency response for Test C.

persons walking randomly in Test C. This is because the response amplitudes depend mainly on the closeness of the excitation frequency to a natural vibration frequency of the structure, in this case associated with its fundamental vibration mode. In Test A, the fourth harmonic of the average walking frequency (4 1.87 Hz = 7.48 Hz) is very close to the fundamental frequency (7.51 Hz) of the uncontrolled structure; this closeness is not attained in Test B (4 1.77 Hz = 7.08 Hz) and in Test C (4 1.74 Hz = 6.96 Hz). 7.1. Efficiency evaluation of the MSDA system (composed of a pair of synchronized TMDs) Table 3 presents the ratios between the controlled and uncontrolled vertical acceleration amplitudes for both the time and frequency (rms) responses filtered to the first vibration mode of the composite floor deck structure. It can be observed that

Table 3 Reductions in rms vertical acceleration amplitudes (av ) in the first vibration mode. Test (Controlled av )/(Uncontrolled av ) Time response (%) A B C 81 58 32 Frequency response (%) 84 70 47

the reductions in vibration amplitudes for the first mode varied from 32% to 81% in time responses and 47% to 84% in frequency responses. The reductions in vibration amplitudes due to the installation of the passive control devices were the greatest for Test A (up to 84% in frequency responses), for which only one person walked in quasi-resonance with the first vibration mode of the structure.

W.D. Varela, R.C. Battista / Engineering Structures 33 (2011) 24852494 Table 4 Subjective classification of the vibrations given by the volunteers. Person Classification Uncontrolled structure 01 02 03 04 05 06 Uncomfortable Uncomfortable Intolerable Intolerable Intolerable Uncomfortable Controlled structure Perceptible Imperceptible Perceptible Uncomfortable Perceptible Perceptible

2493

Table 5 Comparison between rms values of measured accelerations and recommended acceptance acceleration values. Measured acceleration (mm/s2 ) Test A Uncontrolled structure Controlled structure Acceleration values Source ISO 2631-2 [34] criteria Destination Office Conference room Special office SCI [35] criteria General office Busy office AISC [2] criteria Office Shopping mall 161 31 B 77 32 C 72 49

alimit (mm/s2 ) 20 40 20 40 60 35 106

In Test B, where six persons walked side by side in quasisynchronism, the responses were again largely dominated by the first mode, but less so than in Test A. Even so, the reductions in vibration amplitudes were still very large (up to 70% in frequency responses), although smaller than in Test A. Finally, in Test C, where six persons walked randomly (without synchronism in their steps), the reductions in vibration amplitudes were still substantial (up to 47% in frequency responses), although less than in the previous tests. 7.2. Human perception of the vibrations The volunteers for the tests were asked how they felt, and they classified the vibrations in the structure. After test C, one volunteer remained standing at the center of the slab while the other five volunteers walked around. The response amplitudes in the perception tests differ slightly from those in Test C for which all six volunteers walked around. The variation of acceleration amplitudes in these two types of test was found to be only 5%. The volunteers were not informed when the control system was in operation. The options for classification were as follows: (i) imperceptible: the volunteer reported that he could not feel any vibrations; (ii) perceptible: the volunteer reported that he could feel vibrations, but they were not disturbing; (iii) uncomfortable: the volunteer reported that he could feel vibrations, and they were annoying; and (iv) intolerable: the volunteer reported that he could feel vibrations, and they were unbearable. Table 4 summarizes the results. All of the volunteers changed their opinion positively on the vibration level when the TMDs were in full operation as compared to when the TMDs were locked (i.e., out of operation). When the TMDs were locked, half of the volunteers felt uncomfortable, and the other half could not tolerate the vibrations. When the TMDs were released, four of the six volunteers classified the vibrations as only perceptible, one of the other two classified them as imperceptible, and only one still classified them as uncomfortable. Moreover, better results could be achieved with extra units of the same TMD, that is, by increasing the ratio between the TMD mass and the structural modal mass. As human perception is the best way to evaluate vibration attenuation, one can state that the passive control system is an attractive, cost-effective, and successful means to mitigate undesirable vibrations in floor deck structures. Table 5 presents the application of human perception criteria from ISO 2631-2 [34] and acceptance criteria from SCI [35] and AISC [2] to the maximum amplitudes of vertical accelerations in the controlled and uncontrolled structure. The values of amplitudes are related to the time response signals obtained from the experimental tests and frequency band filtered to the first vibration mode of the composite floor deck structure. In Table 5, the prescribed ISO 2631-2 [34] limit values for acceleration related to continuous vibrations were chosen because people walked continuously during the tests. For the uncontrolled structure, it can be noted in Table 5 that the measured acceleration amplitudes are excessive according to all the guidelines when applied to people at rest in conference rooms and offices. For shopping malls, the accelerations measured in Test A are found excessive according to AISC [2].

For the controlled structure, the measured vibration amplitudes (Table 5), although substantially reduced, do not comply fully with the acceptance values related to the guidelines. The acceleration amplitudes measured in Tests A and B are lower than the acceptance values for all destinations other than offices and special offices according, respectively, to ISO 2631-2 [34] and SCI [35] criteria. Conversely, the measured amplitudes in Test C are higher than the acceptance values for all destinations other than busy offices and shopping malls according, respectively, to SCI [35] and AISC [2] criteria. 8. Conclusions The problem of excessive vibration amplitudes produced by humans walking on lightweight large span composite floor deck structures and the effectiveness of simple mechanical synchronized dynamic control devices tuned to a frequency close to the structures fundamental frequency were investigated through experimental measurements and human perception criteria, and also in the light of the acceptance criteria given in design guides. The results show that, for any of the applied human loads, the passive control system provides significant reductions in vibration amplitudes, considerably improving the dynamic performance of the structure and user comfort. It is also worth noting that the mechanical control devices had a total mobile mass equal to only 0.5% of the modal mass related to the first vibration mode of the structure. As such, these low cost and low maintenance passive control mechanical devices can be proposed for the rational design of smart structures of low damping lightweight large span floor decks of buildings intended for residential, commercial, or public use. Acknowledgments We thank the CAPES (The Brazilian Research Council for Enhancement of Faculty Members) and the Instituto COPPE of the Universidade Federal do Rio de Janeiro for financial support. Special thanks go to the volunteers for the walking tests and the technicians of COPPEs laboratory of structures, who made the tests possible. Appendix. Supplementary data Supplementary material related to this article can be found online at doi:10.1016/j.engstruct.2011.04.021. References
[1] Tredgold T. Elementary principles of carpentry, Publisher unknown, 1828. [2] Murray TM, Allen DE, Ungar EE. Floor vibrations due to human activity, Steel Design Guide Series No 11, Pub. No. D811. Chicago (IL, USA): American Institute of Steel Construction; 1997.

2494

W.D. Varela, R.C. Battista / Engineering Structures 33 (2011) 24852494 [21] Magluta C. Passive dynamic systems to absorb structural vibrations, D.Sc. thesis. Brazil: Instituto COPPE, Universidade Federal do Rio de Janeiro; 1993 [in Portuguese]. [22] Shope R, Murray TM. Using tuned mass dampers to eliminate annoying floors vibrations. In: Proceedings of the structures congress XIII of ASCE, 1. 1995. p. 33948. [23] Bachmann H, Ammann WJ, Deischl F, Eisenmann J, Floegl J, Hirsch GH, et al. Vibration problems in structures practical guidelines. Basel (Switzerland): Institut fr Baustatik und Konstruktion, Birkhuser; 1995. [24] Bachmann H, Weber B. Tuned vibration absorbers for lively structures. Struct Eng Int 1995;5(1):316. [25] Rottmann CE. The use of tuned mass dampers to control annoying floor vibrations, M.Sc. Thesis. Blacksburg (USA): Virginia Polytechnic Institute and State University; 1996. [26] Murray TM. Control of floor vibrations state of art report. In: Proceedings of the 15th congress report of international association of bridge and structural engineering. 1996. p. 75162. [27] Battista RC, Varela WD. Remedial measures for vibration of continuous slabs in buildings. In: Proceedings of the XXX Jornadas Sul-Americanas de Engenharia Estrutural. 2002 [in Portuguese]. [28] Varela WD, Battista RC. Controlled reduction of people induced vibrations in large span composite floor deck. In: Proceedings of the 7th EURODYN European congress on structural dynamics. 2008. [29] Battista RC, Varela WD, Santos EF, Correa WL. Measurements and control of people induced vibrations in stadium grandstands. In: Proceedings of the 7th EURODYN European congress on structural dynamics. 2008. [30] Souza JC. Composite spatial structural model for long spans floors, D.Sc. thesis. Rio de Janeiro (Brazil): COPPE/UFRJ; 2002 [in Portuguese]. http://www.coc. ufrj.br/index.php?option=com_content&task=view&id=3686&Itemid=191 [accessed 5.04.11]. [31] Souza JC, Battista RC. Large span composite floor decks. In: Proceedings of the II CICOM, 2nd international congress of steel construction. 2002 [in Portuguese]. [32] Ibrahim SR, Mikulcik EC. A method for the direct identification of vibration parameters from the free response. Shock Vib Bull 1977;44(4):18398. [33] Battista RC, Pfeil MS. Control of wind oscillations of Rio-Niteri bridge, Brazil. Proc ICE Struct Build 2010;163(2):8796. [34] ISO 2631-2: Evaluation of human exposure to whole-body vibration part 2: human exposure to continuous and shock-induced vibrations in buildings (1 to 80 Hz), Switzerland: International Organization for Standardization, 1989. [35] Wyatt TA. Design guide on the vibration of floors, SCI publication 076. Ascot (Berkshire, England): The Steel Construction Institute, 1989.

[3] Lenzen KM. Vibration of steel joistconcrete slabs floors. Eng J Amer Inst Steel Constr 1966;3(3):1336. [4] Varela WD. Theoreticalexperimental model to analyze vibrations induced by people walking on floor slabs of buildings, D.Sc. thesis. Brazil: Instituto COPPE, Universidade Federal do Rio de Janeiro; 2004 [in Portuguese]. http://www.coc.ufrj.br/index.php?option=com_content&task=view& id=3730&Itemid=191 [accessed 5.04.11]. [5] Ji T, Ellis BR. Floor vibration induced by dance-type loads: Verification. Struct Eng 1994;12(3):4550. [6] Allen DE, Swallow JC. Annoying floor vibrations diagnosis and therapy. Sound Vib 1975;3:127. [7] Hernndez-Olivares F, Bollati M, Witoszek B. Static and dynamic behaviour of recycled tyre rubber-filled concrete. Cement Concr Res 2002;32(10): 15871596. [8] Zheng L, Huo S, Yuan Y. Experimental investigation on dynamic properties of rubberised concrete. Constr Build Mater 2008;22(5):93947. [9] Vasconcelos RP. Structural dynamic control by viscoelastic devices, D.Sc. thesis, COPPE/UFRJ, Rio de Janeiro, Brazil, 2003 [in Portuguese]. [10] Vasconcelos RP, Battista RC. Numerical analysis of a composite orthotropic bridge deck with a sandwiched viscoelastic layer. In: Proceedings of the XXXI Jornadas Sul-Americanas de Engenharia Estrutural. 2004 [in Portuguese]. [11] Hanagan LM, Murray TM. Experimental implementation of active control to reduce annoying floor vibrations. Eng J AISC 1998;35(4):1237. [12] Frahm H. Device for damping vibration of bodies, US Patent no 989.958, 1911. [13] Ormondroyd J, Den Hartog JP. The theory of dynamic vibration absorber. Trans Amer Soc Mech Eng 1928;50(7):922. [14] Den Hartog JP. Mechanical vibration. McGraw-Hill Book Co.; 1940. [15] Thornton CH, Cuococ DA, Velivasakis EE. Taming structural vibrations. Civil Eng Mag ASCE 1990;11:579. [16] Allen DE. Building vibrations from human activities. Concr Int: Des Constr 1990;12(6):6673. [17] Gerasch WJ. Reduziering von beim tanzen erzeugten bauwerksschwingungen durch schiwingungstilger, 65, Bauingenieur: Springer-Verlag; 1990. p. 3138. [18] Setareh M, Hanson RD. Tuned mass dampers for balcony vibration control. J Struct Eng ASCE 1992;118(3):63147. [19] Webster AC, Vaicajtis R. Application of tuned mass dampers to control vibrations of composite floors systems. Eng J Amer Inst Steel Constr 1992; 29(3):11624. [20] Battista RC, Magluta C. Spectator-induced vibration of Maracan football stadium. In: Proceedings of the II EURODYN European congress on structural dynamics, 2. 1993. p. 98592.