Dynamic Response and Control of Vibrations in Four Medium Span Steel Footbridges Alejandro Bernabeu Larena, Dr, Bernabeu Ingenieros, School of Architectui ‘nool of Cvil Engineering, Technical University of Madrid Viktor Markelj, Civil Eng ‘Technical University of Madrid (UPM), Spain: °M), Spain; Gregor Cipot, Civil Eng. PONTING 4.0.., Maribor, Slovenia; Jorge Bernabeu Larena, Dr. School of Cvil Engineering, Technical University of Madrid (UPM), Spain. Contact: [email protected] DOE: 10.2749/1208 Abstract ‘The design and performance of four footbridges in Slovenia and Spain were analysed recently to address the dynamic response and control of vibra- tions in medium span (25-50 m) steel footbridges. A comparative analysis of the four footbridges being studied was done, three of them already in service, in order to define the main dynamic char- acteristics and behaviour of medium span steel footbridges. The study includes dynamic load models and anal- ysis considered (response spectra method and time history analysis), as well as the comparison of model predic- tions to experimental results from dynamic load tests. The article shows the sensibility of medium span steel foot- bridges to vertical vibrations induced by pedestrians, and considers the inclusion Of tuned mass dampers (TMD) as an effective and easy to implement system. to improve their dynamic response, reducing the vertical accelerations to meet the required comfort criteria (a) Keywords: dynamic; vibrations; steel footbridges; tuned mass dampers; dynamic load tests. Introduction Over the last years, intense research has taken place into the effects of vibrations of footbridges, which is resulting in the establishment of a series of precise design, analysis and control criteria, providing in particu- lar both methods for considering dynamic loads and criteria to evaluate the degree of comfort of the footbridge." Besides, the recent tendency in footbridge design towards increasingly lightweight and slim structures has led the influence of dynamic actions to take dn greater relevance. In particular, medium span (25-50 m) steel foot bridges present a set of characteris- tics that makes them especially vulnerable to pedestrian induced vertical vibrations. o) (@) In this context, the recent design analysis of four steel footbridges, three of them already in service, is consid- ered in order to address dynamic response and control of vibrations in medium span steel footbridges. The purpose is to establish a comparative analysis and to define the main dynamic characteristics and behaviour of this specific type of footbridges Special emphasis is placed on the pos- sibilities of vibration control by the implementation of tuned mass dampers (TMD), on the dynamic load mode! and analysis considered, and on the comparison of model predictions with experimental results from dynamic load tests Footbridges: Characteristics, Structural Configuration and Static Response Table 1 summarizes the main data and characteristics of the four foot- bridges, designed and constructed between 2004 and 2014, Fig. 1: (a) Stadenci footbridge over the Drava River, Maribor, Slovenia: 2004-2007; (b) Princesa footbridge in Madrid Rio, Madrid, Spain: 2009, (c} Lent-Tabor foosbridge over Drava River, Maribor, Slovenia: 2010; d) Ribja bry footbridge over Lu Ljubljana, Slovenia: 2012-2014, 260. Technical Report Wianica River, Structural Engineering International Nr. 372016Project Location Dates Maximum | Restraint at nultiple span | span (mn) Studenei footbridge over Drava Maribor, Slovenia | 2008-2007 | Multiple span) 42-42-42 River Princesa footbridge ia Madrid Rio | __ Madrid, Spain 2009) Single span 8 Pinned Lent-Tabor footbridge over Drava | _ Maribor, Slovenia 2010 Multiple span | 48-42-42 Pinned River Ribja brv footbridge over Ljubljana 201201 | Slagle span 2485 Pinned fixed Ljubljaniea River Project Main structure [Width (m) | Structural | Slenderness| Structure configuratio height (m) eight (km?) Studenci footbridge over Drava | Triangular steel uss | 32-38 Las 124 ire River Princesa footbridge in Madrid | Double steel beams 45 126 35 Rio Lent-Tabor footbridge over Double steel beams 33 re 136 245 Drava River Ribja bry footbridge over Trapezoidal steel 36 050 50 280" Ljubljaniea River box Only steel structure, without eonerete in compote section, Table 1: Main data and characteristics of the four footbridges considered in the study Studenci footbridge over the Drava River in Maribor, Slovenia, is in fact a reconstruction of an old bridge, first erected in 1885 as a simply supported structure with two intermediate piers ‘on the river, and substituted in 1948 by ‘a new superstructure that retained the intermediate supports. The new Stu- denci footbridge also respects the existing supports that determine the three equal spans of 42 m (Fig. 1). The structural configuration of the foot- bridge comprises a fongitudinal trian- gular steel truss (spine) and a set of secondary transversal beams (ribs). ‘As neither of the banks had enough height to house the structure. the bridge was raised upwards through the walking surface, and therefore is split into two sections, These two lateral sections are joined together at the middle of the footbridge, since the ‘gradual rise of the deck from the sides to the centre allows the structure to be located under the deck in this area, ‘The Princesa footbridge is part of the huge Madrid Rio urban development project along the banks of the Manza- nares River. {tis a 43 m span, simply supported structure with a straight plan view. The structural configuration is simple: a double two steel beams placed on both sides of the footbridge web steel beam, with the beams placed fn the sides of the footbridge, which also serve as handrails. tis a minimal- ist solution in terms of composition, but with a high level of integration in the shape. Key features of the Structural Engineering Internatio footbridge are the gently varying depth in elevation, the vertical stiffeners visible on the exterior sides, and the underside with free edges (Fig. 1). The Princesa footbridge wit its pale green colour is considered a small tribute to the extraordinary Gee- man bridges built after World War IL Not far from the Studenci footbridge in Maribor, the Lent-Tabor footbridge covers a total distance of approximately 120 m between the two banks of the Drava River, and is located at a short dlistance from the Stari Most, or the old bridge of the city. The restrained design of the new footbridge aims to emulate and reinforce the value of the old bridge, which appears to be like a slim band at the lower level, spanning the river in three approximately equal parts, mirroring that of the Stari Most. The structure consists of a classical sys- tem with a steel beam on either side of the deck.” ‘The shape of the beams adapts to create an attractive curved cross-section, where the entire wooden structure forms a single unit, together with the deck andl the lateral railings. Finally, the Ribja brv footbridge over the Ljubljanica River in Ljubljana, Slovenia, is a small, Gansparent and elegant footbridge of minimalist design, allowing unobstructed views along the river. The footbridge is simply sup- ported oa the left bank, while on the right bank the steel deck is fastened in concrete clement, and leans on the existing retaining wall of the Nv. 92016 Ljubljanica River. The resulting ayym- metrical semi-integral structure spans 24.85 m with a thin, trapezoidal steel boxlike form. The steel deck has a maximum height of only 500 mm (slen= demess of 1/50), and is reduced to 250 mm at the edges, giving the struc ture an extremely elegant harp-like appearance. Two of the four footbridges consi ered in this paper are continuous ‘multi-span structures, with very simi- lar span distribution (Studenci_ and Lent-Tabor footbridges), while the other two are single span bridges. Princesa footbridge covers a span of 43 m and is simply supported at both sides, while Ribja brv footbridge is simply supported at one end and rig idly connected at the other, covering a span of 25 m. All the four footbridges are_steel structures. Princesa and Lent-Tabor footbridges comprise steel beams on either side of the deck as the main longitudinal structural element, while Studenci footbridge comprises a trian ular steel truss, and Ribja bry a trap- ezoidal steel box. Figure 2 shows the typical cross-section of the four foot- bridges. Width varies from nearly 3.50 'm to more than 6 m, and the struc- tural height from 0.50 m to 1.75 m, resulting in slenderness between L/24 and LiS0, Table 2 summarizes the considered loads (permanent and live loads), as well as the static response in terms of Technical Report 2612: Typical cross section of the four footbrdges. a) Stuenci; (b} Princesa: (e) Lent-Tabor; (al) Ribja fry. (Units: fos) Project Loads (Nine) "Vertical detection ( Seit-weight + dead | Liveloads | Seltweight + dead | Liveloads | Total deflection loads Toads Studenel footbeidge 23 ao 48 mm (L875) | 115 mm (1/365) [163 mm (1/260) over Drava River Princesa footbridge aa 40 75 mm (L580) | 69 mm (LI630) | 1h mm (LAK) n Madrid Rio - Lent Tabor a6 a0 98 mm (Lit30) | SS mm (LISD) | IN mm (LIS) footbridge over Drava River Ribja brv footbridge 49 a2 90° mm (LAT) | AW mm (SIS) |W mm ALB | over Ljubljanica j River “Total diction without presamber *Construction sages are taken into account (different structural systems), ive load and dead lows on the final structural sytem, selF-weight of the steel structure onthe simple supported beam system, Table 2 Static response (deflection) vertical deflection. It shows that the permanent load values are very simi- lar or even lower than the live loads, depending on the magnitude of pave- ment and finishing loads, and gives an idea of the resulting total vertical deflection in the different structur without considering the initial precamber that varies from L/I80 to L300, Dynamic Behaviour and Control of Vibrations As expected, considering the struc: tural characteristics and. slenderness, the natural frequency falls within the critical range of 1.25-23 Hz (or secondarily 2.54.6 Hz) ia all the four cases: This means the structure is potentially likely to present vibration effects, as this range of frequencies coincides with the excitation caused by pedestrian footsteps, which are found to be at 1.65-2.35 Hz for walk- ers and 1.9-3.5 Hz for joggers. * Figure 3 shows the main natural fre- quency modes of vibration, The two con- tinuous span footbridges, with similar span distribution, present comparable values for the different vertical modes of vibration, being the first four in the critical range or nearby 262 Technical Report As for the single-span_footbridges, Ribja Brv frequency of the first natural vertical mode falls within the critical range, but the frequency of the subsequent vertical modes increases rapidly, moving away from the critical values, It is also interest- ing to notice the modes of vibration fof the Ribja bry footbridge, which clearly shows the singular continuity and rigid connection of one of the supports and the asymmetry of the struetural response. Regarding the horizontal dynamic response, the four footbridges have horizontal frequencies over 5 His this confirms that no interaction with the horizontal dynamic forces induced by pedestrians is expected, the critical range of frequencies for transverse vibrations being between 05 and 12 Table 3 summarizes the vertical and horizontal natural frequencies of the four footbridges. Ic is important to note that steel foot- bridges are very light, with permanent Joad values similar to live loads. This lightness entails a reduction of the modal mass of the system. which has a negative effect on the structure's abil ity to handle dynamic actions. Furthermore, steel structures have Jow critical damping coefficient, which has a negative effect in terms of dynamic response of the structure ‘A.complete dynamic analysis was per- formed at the design stage in each ease, in order to determine the maxi ‘mum vertical acceleration of the struc- ture when subjected to an oscillating load at frequencies in the range of pedestrian footfalls Figure 4 summarizes the results of these analyses, presenting the masimam accelerations obtained in each ease for walking pedestrian load models, and comparing them with the acceleration limits considered® —(Aecelerations below 05 fi: acceptable: Region A: acceptable, Region B: bearable, Region C: unacceptable). Both Princes and Lent-Tabor footbridges present impor- tant a early above the crit ical values, while Stucenei footbridge values range between acveptable and bearable level of comfort. Finally, the accelerations obtained for Ribja_ bry footbridge are much smaller, clearly below critical values. However. since the Ridja brv footbridge frequency (2.76 Ha) is within the characteristic of the excitation eaused by joggers. a further analysis was per- formed to determine accelerations due lerations, Structural Engineering International Nr. 9/2016ta) Mode 1.1.9 hy Mode LL.57 Hz Mode 2.184 He i 4 Mode 4.2.24 1tz Mode 1.2.76 He Mode. 881 He Fig. 3: Naural vertical frequencies. (a) Suudencis ) Lent-Tabor: fe} Rabja bev - Mode 7.3725 He [Froker ‘aural vertical frequencies" (i) Natural horizontal | First mode Second mode Third mode Fourth mode | equeney™ (He) | Studene# footbridge - 1.98 236 322 1SL | over Drava River | Princesa foothridae zal = - 176 in Madrid Rio Lent-Tabor 1s Ta ae 307 3a | oatridge over | Drava River Ribja bev foowidge 176 Ba 7 Ts roa wer Livbhanica, | River The fist four modes are considered Tor the aalural erga GFeQuEREES as They may Fall a the Entel range 12023 Fy and 2546 He (pecond amon “Only: the frst horizontal mode i incat. eing higher than the cra range 1 20130 He Table 4+ Natural frequencies © running pedestrians, Results of this analysis showed that synchronous run- ning of multiple pedestrians resulted in eccelerations greater than 2.50 mus" “hich is considered to rank as unaecept- able behaviour of the structure for this mount of excitation and unacceptable evel of comfort to bridge users.® intervention herefore, required to improve the dynamic response Sh ictural Engineering Interna of structures. Installation of tuned mass dampers (TMDs} was. consid- ered in three of the four foothridges, tuned to attenuate the dynamic response of the laotbridges.*" Only in Studenci footbridge, where the maximum aeceleration calculated (0.78 nvS") was considerably lower than in the other cases. no. special feature to improve the dynantic response was considered. It is ional Nr. 72016 important 10 note that this foot bridge is in service from 2007 and no report of any discomfort to pedestrians has been recorded, Flow- ever, in the context of this study of dynamic performance of middle-span steel footbridges, specific dynamic load tests were performed, as will be described next On the other three structures, TMDs were implemented (on every span in Technical Report 268Fig. 4: Vertical accelerations for watking pedestrian load models. Region Az acceptable, Region Bs beural 1 20 24 28 Pregueney (H2) annecepinble ae RY ie, Region C:1macceptable Project TMD frequency (Hz) Modal mass (kg) [Damping (kNsim) Movement (mm) Princesa footbridge in 205 1265) 62 0 Madrid Rfo Lent-Tabor footbridge over Drava River Lateral span 1 182 2a 282 Middle span 350) 8a Lateral span 2 220) 499) 36 Ribja brv footbridge 2.59 - 92 - over Ljubljaniea River Table 4: Implemented tuned mass damper characteristics Project Maximum vertical accelerations® (mis?) 1 Person 8-15 People Group of people ‘Studenei footbridge over Drava = 0.18 (E372 Ha) - River Princesa footbridge in Madrid = L65 (E261 Ha) - Rio Lent-Tabor footbridge over Drava River 074 (E 307 2) 129 (E157 Ha) = Ribja bre footbridge over Ljubljanica River 0.05 015) oe ‘The masimam vertical acceleration for each dynamic load case ie indicated, indsating the frequency (F) for Which Hw obtained Table 5: Vertical accelerations under walking load models the case of the continuous three span Lent-Tabor footbridge; on the middle of the span in the Princesa footbridge and eecentrivally located, closer to the pinned support, on the Ribja bry foo bridge). Table 4 presents the main characteristies of the TMDs considered in each case, while Tuble 5 and Fig. 5 present the accelerations calculated considering the implementation of TMDs, considerably lower and within the comfort criteria. Vertical aevelera- tions due to walking load cases on Prin- cesa and Lent-Tabor footbridges were reduced below 0.5 avs". while for Ribja brv footbridge vertical accelerations due to running of jumping were also 264 Technical Report considerably reduced to less than 1.0 mvs", which may be considered acceptable for these load cases. The design stage analysis was per- formed by considering the response spectra method at the first stage, ancl therealter the time history analysis, in order to implement dynamic actions due to running pedestrians, and to evaluate and adjust the influence of TMD inclusion.!"" Finally, dynamic measurements con: sidering a group of people walki running or jumping for intentional excitation of the footbridge were per: formed both on Studenci footbridge Structural Ei and Ribja bry footbridge after erec- tion, in order to confirm model predie- tions" (Fig. 6). In the ease of Studenci footbridge it is important to note that at the time the footbridge was constructed, analyses of neither running nor jumping pedestrian Toad cases were performed. And so, while measured accelerations for walk w were within the range of 0 5-1,0mv s*, according to the calculated accelera- tions, for running and jumping the measured! accelerations exceeded maxi- mum aclmissible accelerations. For a running group of 20 people, the maxi mum acceleration —measttred was rmnational Nr. 3/2016(@ 1 12 as 06 aa 1s a 06 emcme cs oo w ‘Vere! aeceleration. 15 people walkie Wit (b)—Avecteration a2 (ausee) ‘Vertical aecteation. 1 people runing, Without TMD Acceleration 4 (see) Vertical aceeration. 4 people sunning. With TMD Ni ke anLesme 77 imo 3 sesneme? omy ” @ Verses aceeration 18 poopie walking, With TMD a) Asselerstion a (assoc) Time (e) Ung =O81L mM gs 025 mi Vertical aceleration 1 people running, With TMD Fielecaion 32 see) Time ) S6timm 4. 095mi Vert aceleration. people runing, Wih TMD Fig, 5: Vertical accelerations with and without tuned mass dampers. (a) Pricesa; (6) Ribja bre 3.50 mvs*, however when the group of. 20 people jumped the acceleration ‘measured was higher than 15 m/e, although in this case the accelerations quickly decreased to values under 110 iS" due to the bridge damping. Fig- ture 7 shows the measured accelerations in side span (higher than in middle span) for walking, running and jumping _roup of people (vertical accelerations are shown in red and horizontal accel- erations are shown in green). ‘The dynamic load tests on Ribja brv were performed considering for both 'TMD blocked and active. The calcu- lated accelerations for walking load cases were well below the maximum comfort levels and no discomfort had been reported or noticed during the first month that the footbridge was in service; however, TMD was not implemented due to a delay in delivery. Experimental load tests therefore focused on running and jumping load cases. Good matching in measurement was obtained in both ceases, confirming the expected effi ciency on implementing TMD_ to reduce vertical accelerations (Fig. 7). Structural Engineering International Nr. 3/2016 Conclusion The analysis. and experimental observation undertaken for four ‘medium span (25 to 50 meters) ste! footbridges recently designed and erected, arrive at the following conclusions that characterize the dynamic response of this type of footbridges. Firstly, the regular structural configu- ration and slendemess of medium span steel footbridges result in natural frequencies that fall well within the critical range of vertical excitation Technical Report 265Fig. 6: Performance of dynamic load tests in Studenci footbridge () [Acsleratonsin X.Y and directions 6 people running, TMD blackod (rd) end active green) 38 or 20 10 ° 10 20 aaa . a CT o 10 Ea 30 o 30 Tine (6) Tine) Aceelerutionsin X.Y ad 7 diretions ts {reopening TMD backed (ed) and ative (eee) ida wo Time(s) oa 5 0 2B 2 ‘Acceltation (mi) 0 Time(s} Fig. 7: Dynamic load tests performance. (a) Studenci footbridge (vertical accelerations in red, horizontal accelerations in green); (0) Ribja brv footbridge (TMD blocked in red, TMD active in green) for different scenarios due to human- induced vibrations (walking, running, jumping), as has been confirmed by induced by pedestrians. Moreover, the vertical accelerations obtained, especially in the case of several walk- system to improve the dynamic response of this type of footbridges, reducing vertical accelerations ing pedestrians or in the case of run- ning and jumping pedestrians, are likely to exceed the maximum admis- sible values usually considered to assure an adequate degree of comfort The inclusion of TMD appears to be an effective and easy-to-implement 266 Technical Report induced by pedestrians in the struc ture, in order to meet the required cri teria of comfort and safety. Time history analysis method was accurate enough to determine the dynamic response of these foot bridges, both with and without TMD. the dynamic load tests. Other simpli fied methods, like the response spectra method, may be used to calet late the maximal vertical acceleration, and to determine if additional mea- sures have to be taken. However they cannot be used for further detailed Structural Engineering International Nr. 3/2016 szsndeeneapeeeameemempmemaimammmaammmamas eeeanalysis, such as to consider walking pedestrians’ dynamic actions or to cal- culate accelerations with active TMD. References [2] SETRA, Technical guide — footbridges Assessment of vibrational Behaviour of foot bridges under pedestian loading, Serviced Etudes Teehmiques des Routeset Autoroute, 2006, (2) FIB. Bulletin 32: guidelines for the design of footbridge. Fédération Internationale du beton (1B), 2005, BB] Heinemeyer C, Bute C, Keil A, Schaich M, et al. Design of lightweight footbridges. for human induced vibrations. Background docs ‘ment in support to implementation, harmoniza tion and further development of the Euroco 3.IRC Scientific and Technical Report, 2000, [4] Markelj V. Footbridge. Studenci over the Drava River in Maribor, Slovenia, Sauce. Eng Ine 21020(4) 483-457, [5] Fernandez Troyano L. Tierra Sobre el ‘Agua. Vision Histrica Universal de’ Los Puentes. Colegio de Ingenieros de Caminos, Canales y Puertos, Madrid, English edition (2003) ~Bridge Engineering: A Global Perspec tive”. ICE Publishing, 1999, [6] Bernabeu A, Bernabe J. Pasarela Sobre el Rio Drava en Maribor (Eslovenia): Analisis y Control de Vibrasiones en Pasarelay Peatonales. V Congreso ACHE Internacional de Estructuras, Barcelona, 2011 (in Spanish). [DI Burges F, Garrido G, Bernabeu A. Lent Tabor footbridge, Maribor, Slovenia. Proceedings of the dk Footbridge International Conference, Wroclaw, Poland, 2011 {8} Studnickova M. The effect of pedestrian trafic on the dynamic behaviour of footbriges Acta Polytech 2006:2(44):47-S1, [9] Bachmann H. Vibration upgrading of gym nasia, dance halls and footbridges. Struct Eng Ine 199232(2):118-124, NO} Bachmann H, Weber B. Tuned vibra absorbers for “lively” steuetures, Sruct Eng In 19955(1)31-36, [11} Caetano F, Conha A. Implementation of a Passive control system ia lively footbridge JABSE Symp Rep 201399(20) 760-767 {12] Keeusinger H. Dynamic Design Soovegies {for Pedestrian and Wind Actions, International Congress on Footbridges: Pais, 2002. [13] Pimemel R, Fernandes H. A. Simpl Formation for Vibration Servceabilty of Foot bridges, International Congress on Footbridges Paris, 2002, [14] Zivanovie S, Pavie A, Ingoltson ET Modeling spatially unrestricted pedestrian tra fic on foorbridges. J Stuct Eng 2010336(10) 1296-1308 [15] Coctano E, Cunha A. Recent perspectives in dynamic testing and monitoring of bridges Siruct Control Health Monit 201320(6):853-87, SEE Data Block Studenci footbridge over the Drava River, Maribor, Slovenia Owner: City of Maribor Designer: Ponting 4.0.0., Maribor (VMarkelj) Contractor Pomgrad d.0.0 Steel (0; 93 Cost (€ 120 million): Service 2007 date! Princesa footbridge in Madrid Rio, Madi, Spain Owner: Designer: Madrid City Hall Burgos y Garrido arquitectos (G. Garrido, F. Burgos); NB3S Ingenieria (I. Jiménez, ‘A-Bernabeu) Contactor: Vins y constructoras, ‘Comse Steet (0) 65 Cost (€ 1.70 million): Service 2009 date. Lent-Tabor footbridge over the Drava River, Maribor, Slovenia Owner: Designer City of Maribor Burgos y Garrido anquitectos (G. Garrido F Burgos); IDOM (A. Bernabeu: J. Bernabeu); Ponting (V. Markel) Contractor ~ 170 Steel (t) Cost (€ 2.00 nilion) Service Unbuitt date: Ribja brv footbridge over Ijubjanica River, Ljubljana, Slovenia Owner: Designer City of Ljubljana Ponting d.o.0, Maribor (G. Cipot, V Markel), Arbitektura doo. (P. Gabrieleic) Contractor: Makro 5 gradaje doo. Koper Steel (t) 26 Cost (€ 0466 million): Service 2014 date el =h-1 4 Dee ener ae A OI ue ean esd Free subscription to quarterly journal SEI Free access to electronic archive of SEI since 1991 GN a anos Cnet ane) ote ners oa iat ete Reduced registration fees for IABSE Conference: eM R mmc ects jpportunity to join activities of National Groups een Latta DSSS cu ee Ronco Es Cle eal Se CRO) Press | Structural Engineering International Nr. 3/2016 Technical Report 267